Amiga Plus 1995 #2/Amiga Plus CD - 1995 - No. 2.iso/internet/faq/englisch/medicalimage-volumevisualizati

Text File | 1995-04-11 | 123.7 KB | 2,594 lines

Archive-name: medical-image-faq/volume-visualization
Posting-Frequency: monthly

med-volviz-faq-95-03
====================

The following is a list of software packages, user's notes and other
information relating to medical volume visualization and imaging that I
have collected over the past from Usenet newsgroups, mailing lists, www
and other places. Many thanks to V.C. Arun Kumar and Lance Ladic for their
lists on 3D visualization software. I have NOT tried all of the packages
mentioned in this list, and therefore cannot attest to the quality of some
of them. Please let me know if I have misquoted someone's posting and if
you wish to make corrections and additions to the faq. Also send me a note
if some of the text is out of date or too far off-topic as I don't want
the faq to grow too much with irrelevant info.

I would be interested to hear other user's experiences concerning 2D and
3D-reconstruction of CT, MRI and US-data.

If you find the faq useful I would be happy to receive a postcard from
your hometown or institution mailed to: Matti Haveri, Oulunsuuntie 122D39,
90220 Oulu, Finland.

WWW-, FTP- gopher- and email-URLs are in a format that newsreaders like
Mac's NewsWatcher <ftp://ftp.acns.nwu.edu/pub/newswatcher/> can pass
automatically to helper-applications.

This file is formatted as ~setext~. For more information on setext, the
structure-enhanced text format send email to <mailto:[e-mail]>.
A file will be returned shortly. Setext viewers for Mac, DOS/Windows and
UNIX can be found at <ftp://ftp.bilkent.edu.tr/pub/Local/setext/>.

This file is also available at:

<ftp://rtfm.mit.edu
/pub/usenet-by-group/news.answers/medical-image-faq/volume-visualization>

<http://www.cis.ohio-state.edu
/hypertext/faq/usenet/medical-image-faq/volume-visualization/faq.html>

<http://www.cs.ruu.nl
/wais/html/na-dir/medical-image-faq/volume-visualization.html>

Matti Haveri <mailto:[e-mail]>
Oulu university central hospital
Department of diagnostic radiology

Changes since 95-02
-------------------

**Software packages:**

IDL updated

XEVA-VisualStudio added

**Some medical sites**

Radiology Teaching Files on the Web added

**DICOM info, software and example image files**

David Clunie's dicom tools updated

Changes since 95-01
-------------------

**Software packages:**

OSIRIS added

NIH-Image DICOM import routine added

**Some medical sites**

Penn State University updated

University Hospital of Geneva - Digital Imaging Unit (OSIRIS & PAPYRUS)
added

General Electric added

RSNA added

CRS4 added

**DICOM info, software and example image files**

ImportACCESS added

Changes since 94-12
-------------------

**Personal info**

I am a proud father of a girl born 9th of December 1994!

**Software packages:**

EVAL3DPET added

SNARK93 added

**Some medical sites**

National Library of Medicine Visible Human Project updated

WWW news.answers archive added

Contents
========

**Software packages**: Progams relating to medical volume visualization
and imaging as well as users's notes.

**Some medical sites**: Some medicine-related WWW-, Gopher-, FTP- and
other Internet-sites worth knowing.

**DICOM info, software and example image files**

**Other interesting FAQs**

Software packages
=================

3DVIEWNIX
---------
Demo (SUN, SGI, and PC) available at (130.91.180.111)
<ftp://mipgsun.mipg.upenn.edu/pub/>. The package sells for $1000 and comes
with the source code, both the Silicon Graphics and SUN versions, and a
couple of data files. You can get some MRI images in ACR-NEMA 2.0 format
at /pub/3DVIEWNIX1.0/DATA.

Contact: Dr. J.K. Udupa, Medical Image Processing Group, University of
Pennsylvania, Department of Radiology, 418 Service Drive - 4th Floor
Blockley Hall, Philadelphia, PA 19104-6021, Phone: [phone], Fax:
(215)-898-9145, <mailto:[e-mail]>.
<http://mipgsun.mipg.upenn.edu>.

3DVIEWNIX is a transportable, very inexpensive software system developed
by the Medical Image Processing Group, Department of Radiology, University
of Pennsylvania, Philadelphia. It has state-of-the-art capabilities for
visualizing, manipulating, and analyzing multidimensional, multimodality
image information. It is designed to run on Unix machines under X-windows.
It uses a data protocol that is a multidimensional generalization of the
ACR-NEMA standards. We have tested it extensively on SGI and Sun
workstations and PCs. Other recipients of 3DVIEWNIX have installed it on a
variety of platforms including IBM RS6000s, HP700s, DEC machines, and
Stardent, all from a single source code version. 3DVIEWNIX has been picked
up as the TOP 10 GRAPHICS SOFTWARE PRODUCT OF THE YEAR by IEEE Computer
Graphics and Applications (January 1994, pp. 87). We charge $1000.00 for
the software which comes with source code and manuals. You can modify and
do whatever else you want as long as it is for your own noncommercial use.

3DVIEWNIX can handle rigid, non-rigid, static, and dynamic objects and
object assemblies. Can handle object information from multiple modalities
and longitudinal acquisitions. Multitudes of visualization, manipulation,
and analysis methods incorporated.

Preprocessing: 1. Volume-of-Interest: *To specify subset of the
n-dimensional (nD) volume image *To specify an
intensity-interval-of-interest for reducing the number of bits. 2.
Interpolation: *To create isotropically sampled data of lower or higher
resolution than input *Many interpolating functions *Interpolation in n
dimensions *Both grey-level and shape-based methods. 3. Filtering: *A
variety of forms of enhancing and smoothing filters *Used for filtering
surfaces, for normal estimation, for interpolation, and volume rendering.
4. Masking: *For assisting segmentation *Quick operation using "paint
brushes". 5. Thresholding: *Multiple intervals can be specified
*Iso-surface generation at any resolution. 6. Segmentation: *2-feature
cluster partitioning *Quick gesture-controlled (user-guided) boundary
segmentation. 7. Classification: *1-feature multiple material
classification for opacity assignment *2-feature multiple material
classification for opacity assignment. 8. Boundary Formation: *Connected,
oriented, closed 3D surfaces are formed *Surfaces may have any resolution.
9. Image Algebra: *Image addition, subtraction, logical operations.

Visualization: 1. Slice: *Sophisticated form of slice display *Multiple
input volumes of any dimensionality can be handled simultaneously
*Multiple color maps *Static montage viewing and dynamic cine viewing of
slices *Arbitrary magnification. 2. Reslice: *Guided by 3D display
*Reslicing through multidimensional volumes. 3. Surface Rendering:
*Multitudes of methods *Multiple objects with translucency and color
*Based on the notion of a structure system: A structure system may be a
collection of static objects, dynamic rigid objets, dynamic non-rigid
objects or any of these coming from multiple modalities *Structure systems
are visualized in their natural form, e.g., a beating heart is displayed
in that manner *Viewing properties of objects can be changed
independently. 4. Volume Rendering: *A new very fast method called shell
rendering *Interactive rendering *Interactive opacity and color
modification *Interactive measurement of fuzzy surfaces.

Manipulation: *One of the most sophisticated set of operations in
3DVIEWNIX *A variety of complex operations including cut away, reflect,
separate, move, surface marking, measure, animation *Complex surgical
procedures can be simulated.

Analysis: 1. Measurement: *A variety of inter and intra structure
morphometry *A variety of image intensity-based measurements such as
density profile, time density curves, region-of-interest statistics and
their variation with time. 2. Registration: *Based on matching homologous
features - points, curves, entire surfaces *For merging information from
multiple modalities *For motion description and analysis. 3. Motion
Analysis: *Rigid object assemblies *Animation of motion and its
quantification *Comparison of motion of two assemblies of objects such as
two joints *Relationship between moving surfaces.

Ongoing Work: *Fuzzy connected component object segmentation *A variety of
user-steered quick segmentation strategies: live-wire, live-band,
live-region methods *Fast volume rendering of fuzzy structure assemblies
with digital perspective *Manipulation of shells (fuzzy objects) and shell
algebra *Registration of shells (fuzzy objects) and their motion analysis
*Portable system integration.

ANALYZE
-------
Provides an environment for the interactive visualisation and manipulation
of 2-D, 3-D and 4-D biomedical images. An integrated set of tools is
provided to allow data to be interrogated in both two and three
dimensions. Three dimensional rendering tools are integrated with two
dimensional orthogonal displays to allow real time reconstruction of
conventional 2D. ANALYZE provides all the tools, including image
registration, to truely support multi-modal image analysis. Tissue
characterisation from multiple MRI, CT X-ray, and Nuclear medicine data is
available as an interactive tool. Filtration program allow data
preconditioning from statistical spatial filtering to minimise noise, and
advanced 3D frequency domain deconvolution of the point spread function of
a confocal microscopy system. ANALYZE has been deliberately developed to
run on standard computer hardware. This allows for maximum value to be
gained from the continual development of computer hardware. All supported
customers receive unlimited hotline support and and an annual two day
training course. ANALYZE is available to clinical and academic users for
$16,000. Annual support contracts are $2,000. Contact CNSoftware for
further details and evaluation.

Computers supported: Sun Sparc stations (Solaris 1 & Solaris 2); Silicon
Graphics; HP 9000/700 series; DEC station 5000; DEC alpha; IBM RS 6000;
Apple Mac Quadra. X-Windows supported on all platforms. 24 bit colour
supported where available.

Interactive 2-D image display: Display of multiple images with variable
size control; Mouse driven intensity windowing; Rapid generation of
orthogonal images from 3-D volumes; Display of 3-D volume as a cube with
control of size, intensity, range, angle-of-view and interactive
dissections along orthogonal planes; Generation and display of arbitrary
oblique planar images through 3-D volumes; Interactive generation of
"curved" images and/or radial image sections through images traced on
orthogonal images; Rapid display of images in cine movie loops.

3-D Image segmentation: Semi-automatic segmentation using advanced
morphology operations; Manual editing and automatic connection/deletion of
multiple objects using region growing; 3D image editing and object
definition; Multi-modal image classification and object definition.

Advanced 3-D image manipulation: Volume rendering using ray casting to
display 3-D images from volumetric data. A complete suite of facilities is
provided: Depth, depth gradient, grey scale and gray scale gradient shaded
surfaces; Maximum intensity projection with optional depth weighting.
Variable illumination and angle of view; Dynamic viewpoint manipulation;
Transparency for overlying surface structures; 3D interaction between
objects and orthogonal 2D slices; Multiple rendering parameters on
different regions of the same display; Combined display of multiple
segmented objects using different rendering parameters and colours.

Interactive surface labelling: Surface rendering for display of shaded
surfaces from contours extracted from segmented image data; Surface
smoothing and enhancement based on local neighbour characteristics within
the data; Display and output of surface contour profiles; ASCII file
output of surface normals for export to CAD/CAM or other design or
prosthetic applications.

Multi-modal Image Analysis: Geometric image registration across multiple
modalities using object surfaces or point files. Multi-modal image
analysis and segmentation. Fused image generation and display. Cross modal
object display - Bone from CT X-ray with soft tissue MRI.

Image & Data Manipulations: Linear combinations of images using algebraic
operators; Pseudo transparent addition of multimodal data; Spatial and
frequency domain image processing using standard and user defined filter
functions. Histogram operations. Manual object segmentation using
thresholding, tracing and erasing. Semi-automated, interactive boundary
detection for object segmentation. Automatic edge contour extraction. 2-D
and 3-D math morphology operators. 2-D and 3-D image transformation
compression using wavelets.

Image measurement: Plotting of line and trace profiles including 3-D
tracing. Region growing and spline region definitions. 2D and 3D region of
interest definition. Selection and automatic sampling of regions of
interest with image parameters. Interactive regional volume calculation.
Regional shape and texture analysis. Data plotting and statistical
analysis. 2-D and 3-D shape measurement tools. Multi spectral image
classification tools for multimodal data characterisation.

Operators toolbox: Escape and return to UNIX shell to run user developed
programs. Macro facility to record and rerun display and analysis
sessions. Magnifying glass for magnification of different areas of the
screen at different sizes. Full control of colour palette. Tex generation.

Image review: B/W & Colour postscript printer support. Multi panel cine
movie. Save facility for 24 bit RGB images for review or advanced
printing.

Software development: Support of developments of Analyze program
extensions simple user defined menu builder. Within the Analyze
documentation you will receive a sample 'C' code program to help you
develop your own utilities. Access to Analyze shared memory from external
program and interface building tools.

Data types and structures: As an inherently modality independent
environment Analyze naturally allows the comparison and the fusion of data
collected from different sites or scanners, or from different modalities.
CNSoftware can assist you with porting data into Analyze by developing
additional file conversion utilities. File import facility for: IGE Signa,
Advantage windows, 9000 & 9800; Siemens Magnatom; ACR/NEMA; Interfile;
Papyrus; TIFF included. Within the Analyze documentation you will receive
a sample 'C' code program to develop your own import file utilities.
Analyze is fully compatible, for file import, with the widely accepted
ACR-NEMA file structure and the Papyrus format. These have been adopted by
many scanner manufactures including IGE; Phillips and Siemens and which
the majority of scanners will support as a file downloading format.
Analyze will support a wide range of data resolutions including binary; 8
bit; 16 bit; 32 bit; and 64 bit data. All measurements are made at full
data resolution irrespective of the display resolution which may be
adopted. Import/export from standard image formats, TIFF; Sunraster; PCX;
GIF; PPM etc.

Confidence: Developed at the MAYO FOUNDATION'S Biomedical Imaging
Resource, Analyze allows you to benefit from more than 15 years experience
in visualisation of biomedical data and to have a software team of 15
members behind your imaging applications.

Value: The independent workstation approach offered by Analyze allows for
more productive use of scanner systems by allowing data analysis and
research to be carried out without requiring access to the scanner
console. Where not all modalities are present at a single site Analyze
allows the maximum value from externally contracted studies by allowing
referring clinicians full access to the data on their patients. Staff can
also develop and maintain their skills in the full range of data
modalities even when these are not directly available.

Training: All new purchasers receive a two day training course.

Research and Collaboration: Training of new staff with new data modalities
and new imaging approaches such as 3-D, is a realistic proposition with
Analyze. For senior clinicians the independence offered by Analyze means
that data collected over the years of clinical study and research can move
with the clinician from hospital to hospital. Collaboration between
clinicians and scientists at different institutions is greatly facilitated
by the harmonisation of data display which can be achieved by using a
standard resolution and colour palette in Analyze.

Upgradability: Analyze is supplied as an upgradable product. Entry into
the upgrade program, at the time of purchase, is available for an annual
payment of $2000. Including a further 2 day training course. Analyze is
available to Clinical and Scientific sites for $16,000.

For further information call CNSoftware: In Europe: Phone [phone],
FAX [phone], <mailto:[e-mail]>, CNSoftware Ltd, The Old
Post Office, Worthing Road, Southwater, W.Sussex, RH13 7DT. In North
America: Phone [phone], FAX [phone], email
<mailto:[e-mail]>, CNSoftware Inc, 201 1st Avenue SW, Rochester,
MN 55902.

AVS
---
Commercial package from Advanced Visual Systems, Inc. AVS is a
visualization application software and development environment. AVS
accepts data and attempts to create a visual display of the data in a
variety of forms using different visualization techniques. AVS is
structured around their concept of a module. A module is an independent
computing element (C or FORTRAN) which is represented by a rectangular
icon on the AVS screen. AVS comes with 110 modules, and the International
AVS Center provides access to a much larger set of modules contributed by
the AVS user community. A range of data input, filter, mapper and data
output modules are also included in AVS. Filters transform data into data,
e.g. contrast stretch or edge detect. Mappers transform data into
geometry, e.g isosurface or arbitrary slice. And data output modules write
data to files, send data to peripheral devices, or render data, e.g
displaying geometry, images and volumes on the screen.

Convex, DEC, Hewlett-Packard, IBM, Sun, Wavetracer.

<http://sslab.colorado.edu:2222/projects/AVS_toc.html>.
<ftp://avs.ncsc.org/>. Usenet: comp.graphics.avs. Contact: Advanced Visual
Systems Inc, 300 Fifth Avenue, Waltham, MA 02154, USA, Tel:
1-800-428-7001, 617-890-4300, Fax: [phone], <mailto:[e-mail]>,
<mailto:[e-mail]>.

Biological Detection Systems
----------------------------
15200 Omega Drive #105, Rockville, MD 20850, Phone: 800-BDS-7706, Fax:
301-990-8391. Macintosh.

>Pete Clinch: They are a company that has been around in one form or
another for the last 5 years or so. It was formed to capitalize on
technology developed at Carnegie-Mellon's Center for Fluorescence Studies.
The package they sell is a turnkey system that comes with a computer. The
main product is the software, however, and includes such niceties as
removal of out-of-focal-plane light to de-blur images and 3-D
reconstruction of a spatial series of images. Their software is at least
$10,000 for the basic package, and it can be much more than that depending
on how many of their add-ons you buy.

Bioquant
--------
Low-end 3D reconstruction and quantitative histochemistry system.
Platforms: PC. Contact: R&M Biometrics, 5611 Ohio Ave, Nashville, TN
37209, Phone: [phone].

CT programs by Malcolm Slaney
-----------------------------
<ftp://ftp.apple.com/pub/malcolm/ct.tar>

>Doug Merritt: This is a package of code that accompanies the book
"Principles of Computerized Tomographic Imaging" by A. C. Kak and Malcolm
Slaney (IEEE Press, 1988). It has nothing to do with my work at Apple.
Programs included: back - Back Projection, bgr - Insert Black Cross Hatch
on the Comtal, cen - Find the Center of CRC Scanner Data, disn - Quantize
an N x N picture, filt - Filter Projection Data, fmm - Find Minimum and
Maximum of Data, g128 - Generate 128 x 128 Picture from Quadrants, gen -
Generate Simulated Data for CT Scanners, hf - Homorphic Filter Waveforms,
hist - Find Histogram of Eight Bit Binary Data, kakman - Print Help Files
for CRC Tomography Software, median - One Dimensional Median Filter, merge
- Add Two 128 x 128 Images together, path - Generate multipath data,
pdsname - Extract PDS Name Information, radon - Plot the Radon Space of
the Scans, scan - Massage Raw Data from the CRC Scanner, sim - Simulate
Ellipse Field Images, tof - Calculate Time of Flight vs. Threshold Value.,
tv - Extract True Values from Comtal Images, window - Convolve an image
with a square window (averaging).

DIP Station
-----------
Macintosh. Contact: <mailto:[e-mail]>.

>Pete Clinch: I'm pretty happy with a package called DipStation. I think
you can also make user code modules that are basically written in c.

Dr Razz
-------
Dr Razz is a 16 bit image display and analysis program for Macintosh color
computers. The program has been optimized for display of radiographic CT
and MRI images, although any 16 bit image stored in a raster file format
(with or without a header) can in principle be viewed. Features include
near real-time window width and window level adjustments on the full 16
bit image data on standard Macintosh graphic hardware. Images can be
viewed individually, or a series of images (eg, a CT or MRI exam) can be
viewed in an image stack. Most non-compressed CT and MRI images can be
opened automatically, without entering any image parameters. In the 'Auto'
open mode, the program attempts to automatically determine image type (CT
vs. MRI), presence of a header and byte order (little endian vs. big
endian). However, a 'Custom' open mode allows complete adjustment of these
and other parameters. Images created with the General Electric 'ximg'
image extraction tool can be opened directly, even if compressed. The
window width and window level setting can be interactively changed via the
window/level control, or by the arrow keys. Most of the image processing
and image analysis tools are not yet implemented. Images can be saved as
16 bit raster files, or 8 bit grayscale PICTs. TIFF and 8 bit raster
formats should be included by the first official release.

The application supports the core AppleEvents (except for printing, which
is not yet implemented) and stationary pad documents.

System 7.x and a color Macintosh with a 68020 or greater CPU are required.
A math co-processor is NOT required. A Power Macintosh version will be
released after the first non-beta version is posted for 680x0 Macintoshes.

Dr Razz is a "freeware" application. You are free to distribute this
program for non-commercial use. Please include all documentation that came
with the program. However, the copyright is retained by the author,
Thurman Gillespy III.

I am interested in supporting as many file formats as possible. Please
contact me if you have specific file format information. C language header
files are especially appreciated. I am also interested in collecting as
many examples of different image file formats as possible for a test
suite.

Bugs reports, and any general comments about the program should be sent to
<mailto:[e-mail]>. Inquiries to the author should be sent to
<mailto:[e-mail]>. Thurman Gillespy III, Department of
Radiology, SB-05, University of Washington, Seattle, WA 98195, voice:
206-543-3320, FAX: [phone].

<ftp://ftp.u.washington.edu/pub/user-supported/razz/>

EutecticSSRS
------------
Low-end 3D reconstruction, mapping, and analysis system. Contour-based
using a digitizing tablet. Platform: PC. Contact: Eutectic Electronics,
Inc, 8608 Jersey Court, Raleigh, NC 27613, Phone: [phone] (also:
919-782-3000).

EVAL3DPET
---------
EVAL3DPET - Programs for the Evaluation of 3D PET Reconstruction
Algorithms are available from the Medical image processing group,
Department of radiology - University of Pennsylvania.

EVAL3DPET is a set of programs designed to statistically evaluate 3D PET
reconstruction algorithms. It comprises of tools for 3D phantom and
projection data generation, evaluation and statistical comparison, and it
includes some fully 3D reconstruction algorithms.

The EVAL3DPET programming system is designed to be capable of:

Generating phantom and projection data based on a realistic 3D PET scanner
model. The phantoms are random samples from a statistically described
ensemble of 3D images with 69 ellipsoid features (cold, normal and hot)
ranging from small (4 mm) to large (40 mm). The projection data generation
takes into account detector field-of-view blurring and a realistic 3D PET
noise model.

Evaluating for structural accuracy, hot spot detectability and cold spot
detectability. The evaluation program also calculates a training
figure-of-merit, that can be used for optimizing reconstruction
techniques.

Statistically comparing the efficacy of a pair of reconstruction
techniques using the t-test.

Included, as examples, are the ART and EM reconstruction techniques, both
implemented using traditional voxels and also using the so-called "blob"
basis functions. Our implementations of ART use a special data-access
ordering (of projection rays) to achieve fast convergence. The EM
algorithms support both attenuated and non-attenuated projection data.

EVAL3DPET will be made available to all who request it at the cost of
reproduction and mailing of the C source code (on a UNIX tar tape) and the
manual. The programs were developed using the C language (K&R), under a
UNIX operating system, and they have been tested on SPARCstations (SUN)
and Silicon Graphics machines. The software and the manual may also be
received via ftp (in which case we will require a login ID and a
password). There is a charge of $150.00 (checks only) for providing this
service. (For overseas mailing add another $50.00 if air mail delivery is
required.) Please make the check payable (in US currency) to RADIOLOGY
ASSOCIATES and send it with your order to:

Ms. Mary Blue, Medical Image Processing Group, Department of Radiology -
University of Pennsylvania, Blockley Hall, 4th Floor, 418 Service Drive,
Philadelphia, PA 19104--6021, Tel.:[phone], fax:[phone],
e-mail: <mailto:[e-mail]>.

By purchasing the EVAL3DPET software, the recipient agrees to abide by the
following terms:

1. EVAL3DPET shall not be redistributed in any way.
2. EVAL3DPET is not a patient-care tool and it is not approved by the
United States Food and Drug Administration.
3. While every effort has been to correct all known bugs, EVAL3DPET is
provided "as is" with no warranty whatsoever. As such, the recipient
agrees not to hold the authors responsible for any problems they may
encounter with the software.
4. The recipient agrees to purchase EVAL3DPET with the explicit knowledge
that the authors do not offer technical support.

FAST
----
Currently under development by members of the Numerical Aerodynamics
Simulation (NAS) Division at NASA Ames Research Center, Moffett Field, CA
94035-1000. It is a software environment for analyzing Computational Fluid
Dynamics data. FAST consists of a collection of separate programs
(modules) that run simultaneously and allow the user to examine the
results of numerical simulations by loading data files, performing
calculations on the data, visualizing the results of these calculations,
and constructing scenes of 3D graphical objects that may be animated and
recorded. SGI.

<http://www.nas.nasa.gov/FAST/fast.html>. User Guide $72, Source code for
commercial customers $2000, for educational institutions $200. Contact
COSMIC, phone [phone], fax [phone],
<mailto:[e-mail]>. To join the FAST user group and
receive tips and important announcements about FAST, send your email
address to: <mailto:[e-mail]>. Email questions,
comments and suggestions to: <mailto:[e-mail]>.

GVLware
-------
Bob - An interactive volume renderer for the SGI.

Raz - A disk based movie player for the SGI.

Icol - Motif color editor.

Contact: <mailto:[e-mail]>. Free. The Army High Performance
Computing Research Center (AHPCRC) has been developing a set of tools to
work with large time dependent 2D and 3D data sets. In the Graphics and
Visualization Lab (GVL) we are using these tools along side standard
packages, such as SGI Explorer and the Utah Raster Toolkit, to render 3D
volumes and create digital movies. A couple of the more general purpose
programs have been bundled into a package called "GVLware".

The most interesting program is probably Bob, an interactive volume
renderer for the SGI. Some Bob features: Motif interface, SGI GL
rendering. Renders 64 cubed data set in 0.1 to 1.0 seconds on a VGX. Alpha
Compositing and Maximum Value rendering in perspective (only Maximum Value
rendering on Personal Iris). Data must be a "Brick of Bytes" on a
regularly spaced grid. Animation, subvolumes, subsampling, stereo. Raz
streams raster images from disk to an SGI screen enabling movies larger
than memory to be played. Icol is a color map editor that works with Bob
and Raz. Source and pre-built binaries for IRIX 4.0.5 are included.

<ftp://ftp.arc.umn.edu/pub/gvl.tar.Z>
To use GVLware:
mkdir gvl ; cd gvl
zcat gvl.tar.Z | tar xvf -
more README

IAP
---
Imaging Applications Platform is a commercial package for medical and
scientific visualization. It does volume rendering, binary surface
rendering, multiplanar reformatting, image manipulation, cine sequencing,
intermixes geometry and text with images and provides measurement and
coordinate transform abilities. It can provide hardcopy on most medical
film printers, image database functionality and interconnection to most
medical (CT/MRI/etc) scanners. It is client/server based and provides an
object oriented interface. It runs on most high performance workstations
and takes full advantage of parallelism where it is available. It is
robust, efficient and will be submitted for FDA approval for use in
medical applications. Cost: in the $5K range.

Available from: ISG Technologies, 6509 Airport Road, Mississauga, Ontario,
Canada, L4V-1S7, [phone].

IBM Data Explorer
-----------------
IBM, HP, Sun, SGI, DG. Contact: <mailto:[e-mail]>.

>Keith Sams: Data Explorer is used in the following commercial industries:
CFD, Earth Sciences, Environmental Modeling, Power Generation (ie
Utilities), Financial Modeling, Petroleum Exploration, Virtual Reality ,
Computer Graphics Education, Medical Imaging, ECAD (Electronic component
design), Site Remediation, Chemistry/Molecular Modeling. To get another
view of how Data Explorer is being used you might want to look at
<http://www.tc.cornell.edu/DX/dx.html> and
<http://www-i.almaden.ibm.com/dx/>. There are some great application
examples there and some example mpeg animations done by students in the
computer graphics program at Cornell.

IDL
---
Interactive Data Language is a package for the interactive reduction,
analysis, and visualization of scientific data and images. IDL integrates
a responsive array oriented language with numerous data analysis methods
and an extensive variety of two and three dimensional displays into a
powerful tool for researchers. IDL supports an extensive data import
capability, publication quality hard copy output, and user-defined
Windows, Macintosh, or Motif graphical user interfaces. IDL is useful in
physics, astronomy, image and signal processing, mapping, medical imaging,
statistics, and other technical disciplines requiring visualization of
large amounts of data. Environments: Macintosh, Unix, VMS and Windows.
Cost: $1500 to $3750, Educational and quantity discounts available.

Contact: Research Systems Inc., 2995 Wilderness Place, Suite 203, Boulder,
CO 80301, USA, Phone: [phone], FAX: [phone],
<mailto:[e-mail]>.
<http://sslab.colorado.edu:2222/projects/IDL/idl_ssl_home.html>. Demos:
<ftp://ftp.rsinc.com/pub/idl/> (all OSs),
<ftp://boulder.colorado.edu/pub/rsi/idl/> (all OSs),
<ftp://ftp.Germany.EU.net/shop/CreaSo/IDL/>. Usenet: comp.lang.idl-pvwave.

>Joe Biegel: [about commercial packages] ...I'd recommend looking into IDL
from Research Systems in Boulder CO. It does many things (including volume
rendering now) real well. It is also very programmable & extensible
(unlike some more turnkey packages).

>Melissa A. Hines: I have compiled a brief summary of what seems to be the
net-consensus. The general consensus is that IDL is an excellent package
_if_ you realize what you are getting. From the IDL Readme file: "IDL,
Interactive Data analysis Language, is a complete package for the
interactive reduction, analysis, and visualization of scientific data and
images." General notes:
o IDL is an interpreted language that has its roots in the PDP world. In
other words, this program predates the Mac (and mice) by many years.
Nevertheless, it is a very impressive package.
o IDL makes extensive use of a command line interface -- more than any
other Mac program in common use. You can program a GUI using "widgets."
From the responses to my post, it would seem that many people DO NOT do
this; however, the people who have attempted it say that it is not too
difficult to use.
o IDL runs on many different platforms -- workstations, mainframes and
personal computers. A number of replies indicate that the Mac version has
a higher density of bugs. Apparently, the manufacturers are aware of this
and trying to recruit a true Mac programmer.
o Technical support for IDL is excellent, but you have to pay for it
(approx. $200/yr.) On the other hand, there is a newsgroup devoted to IDL
(and its cousin, pvwave) -- comp.lang.idl-pvwave. There are also a number
of repositories of IDL code scattered about the net. See the IDL FAQ on
comp.lang.idl-pvwave for more info.
o IDL costs big bucks (on the order of $1500 I have heard). On the other
hand, there is a demo version available.

>Amara Graps: IDL is a vector-based language that makes it easy to
manipulate arrays and matrices. If A and B are arrays, then you can
multiply them together without a FOR loop, letting IDL worry about
accessing each index: C = A*B. However FOR loops are available if you need
them. (Because it is an interpreted language, some actions slow down the
computation, and using FOR loops is a biggie. I've done testing comparing
IDL speed to Fortran in various actions, and IDL was as fast as a Fortran
program in many cases.)
The scientific functions and procedures that come with IDL are often all
that scientists need, at least when you first start out. And if you need
to do some computation where a function doesn't exist, users over the
years have contributed a lot of routines to various archives all over the
world (the two at John Hopkins and at Goddard are especially good). The
language, for the most part is "open", i.e. you can see the text of any
particular procedure or function, in case you doubt the technique, or want
to modify it. Some functions and procedures are black-box, intrinsic
functions or procedures, but not nearly as many as Matlab (see below) are.
Keep in mind that you are in a scientific data analysis environment whose
roots are *not* in the Mac world, so you don't have nice user-interface
items to lead you through every step of the data analysis process, like
you might have with Igor Pro (I have't used Igor Pro, just Igor 1.2, and
didn't like it that much.)
Matlab? It's a similar scientific data-analysis environment, with
capabililties to build GUI programs, and it costs about the same: ~$1500.
Since I've spent the last 4 months converting Matlab wavelets code to IDL,
let me tell you some of the differences. IDL is more of a true programming
language. Matlab has scripts and functions and no way to explicitly type a
variable. IDL has programs, procedures, and functions and a language
syntax sort of like a cross between Fortran, Pascal, and APL. Matlab's
syntax is much more compact than IDL's. (For example: x = transpose(y) in
IDL is x=y' in Matlab.) Matlab has many more built-in, intrinsic functions
than IDL. MathWorks, the company that makes MatLab, has a thriving
business selling Toolkits, such as a Signal Processing Toolkit, which are
libraries of more intrinsic functions, for a fairly steep cost (I think).
>>First, IMHO Khoros and either IDL and MATLAB are complimentary rather
than duplicative as I first thought. >Yes, Yes, Yes! RSI, makers of IDL
realized this too, and so they have built a set of routines to link IDL
with AVS, a boxes-and-wires type of visualization system similar to
Khoros. I've seen and heard alot about Khoros- it seems to have far more
useful data-analysis routines than SGI's Explorer, IBM's Data Explorer or
AVS.

>Mike Schienle: Just to follow-up to a much earlier suggestion of mine
regarding using IDL for the PowerMac. I tested the speed of the built-in
demo on a PMac 7100 (66 MHz) vs. a Sun SPARCStation 10 (50 MHz). The Mac
had 24 MB RAM. The Sun had 256 MB RAM. The IDL demo does quite a bit of
FFT, blurring, line plotting, etc. In short, I feel it is a decent test of
Int, FP and video speed for a system. Anyway, a single pass on the PMac
took six minutes. The same demo running on the Sun SS-10 took 4.5 minutes.
Roughly, a 6:1 price ratio and a 4:3 performance ratio between the Sun and
PMac (for running IDL only). The demo was less than half-way done on a Mac
IIci when the 10 minute demo license expired.

>John C. Schultz: Khoros provides one of the best sets of image processing
algorithms I have seen plus adequate 2 and 3D visualization tools and it
seems to have a large "installed" base. It is also easy to use with a neat
graphical "data flow" interface. It is lacking in interactive quantative
features (interactive ROI, displaying numeric values, profiles, etc) and
has limited data analysis features.
IDL has very good interactive image display features and has good data
analysis tools though not nearly so many high level image processing
routines as Khoros or as many data analysis routines as MATLAB. The
command line syntax is REALLY bad IMHO.
I have not yet found out about MATLAB's interactive display features (part
of an on-going evaluation) but MATLAB has zillions of data analysis
algorithms with many neat and useful ones in the toolboxes. The image
processing toolbox is very bad IMHO since images are stored as DOUBLE (64
bit precision) in the range [0.0, 1.0]. As with IDL the command line
syntax is bad news IMHO. The user base is huge however meaning that lots
of neat program (M-files I guess) are available.
Both IDL and MATLAB run on Windoze, Mac, Unix and probably NT which pretty
much covers everyone's favorite OS-tipple. Khoros is only Unix/X but does
run with Linux.

>Richard Olsen: If you want to do image analysis, IDL would have been an
automatic choice over MATLAB until recent times. Since IDL has firm roots
in the imaging world, vs signal processing, it is very adept at
manipulating images (or any array of information). Now MATLAB has an image
processing toolbox, and neural net toolbox that help balance out the
origins of the packages. IDL can read virtually any data structure known
to man, using existing io-procedures. You can also set up your own. My
students had trouble using MATLAB to read oddly formatted data sets...

>Chris Ruckman: The topic of comparing Matlab to IDL comes up often in
comp.soft-sys.matlab, although it's not in the Matlab FAQ. If you want
more detail, you might post there to request a recap of the last go-round.

Image Pro
---------
From Media Cybernetics ($3,500).

ImageSpace
----------
Software environment for confocal imaging. Integrated acquisition,
processing, and visualization for 3D datasets. Platforms: SGI. Contact:
Molecular Dynamics, 880 East Arques Avenue, Sunnyvale, CA 94086.

ImageVolumes
------------
Interactive image processing, contour editing, 3D reconstruction and
volumetric analysis for confocal, EM, X-ray tomography, and MRI.
Platforms: SGI. Price: $4,500, software maintenance is $950/yr. (10/94).
Contact: Minnesota Datametrics Corporation, 1000 Ingerson Road, St. Paul,
MN 55126-8146, Phone: [phone], Fax: [phone],
<mailto:[e-mail]>. Demo versions of ImageVolumes are available at
the Gopher site gopher.mndata.com 2074 (look in the US/Minnesota listings
under Minnesota Regional Network).

SOFTWARE DESCRIPTION

The ImageVolumes system for the Silicon Graphics IRIS workstation provides
a complete software environment for three-dimensional (3D) reconstruction,
visualization and quantification of volumetric data. Input can be
digitized gray scale images, or two-dimensional (2D) graphics data that
describe contours and points. You can interactively process either form of
data to produce sophisticated 3D shaded surface models. An image
processing module, Image, lets you enhance and analyze serial section gray
scale images using several different classes of functions, including
radiometric, filtering, algebraic, geometric and morphologic. A 3D
graphical editing module, ContourEdit, lets you edit and align serial
section contours and point data taken from digitizing tablets or
microscope stage digitizers such as the MD2 Microscope Digitizer from
Minnesota Datametrics. The 3D display and analysis modules, Display and
Metrics, make full use of the visualization features of the IRIS
workstations including their surface materials and light source modeling
capabilities and their fast hidden surface removal and polygon rendering.
Most importantly, you can make measurements on 3D models such as
distances, numbers of objects, surface areas and volumes. In addition,
ImageVolumes now includes new advanced analytical tools for quantification
and classification of 3D models, such as the distance field and 3D model
intersection tool, DField.

OVERVIEW

You control actions in ImageVolumes using an interactive, screen oriented
DataFlow Manager. By simply clicking the mouse on one or more icons you
define the sequence of operations on your data and the results to be
displayed and stored. At each step of the process the DataFlow Manager
assists you in the selection of appropriate input and output data files.
The major programs within ImageVolumes are:

Image - An interactive image processor that operates on digitized gray
scale images.

ContourEdit - A screen oriented 3D database editor.

Cubes - A polygon and voxel generator that processes your serial section
data and stores a 3D geometry database of surface polygons and voxel
densities.

Display - Displays your reconstruction with user defined surface materials
properties, Phong shading and multiple light sources.

Metrics - A volumetric analysis program for measuring sizes and numbers of
objects and their surface areas and volumes.

DField - Computes distance fields of 3D surface models and quantifies
intersections between 3D surface models. Distance fields can also be used
to interpolate iso-surfaces between two 3D models.

Image Processing

Image is an interactive 2D image display and processing program. Images
can be displayed singly, as movie loops or as a mosaic. You can choose
from among radiometric, algebraic, geometric, filtering , morphological
and graphical overlay functions. The program supports region-of-interest
processing and can be run using scripts or macros that are learned by the
software during your processing operations. Major image processing
categories in Image include:

Radiometric operations - Gray level scaling, histogram normalization,
histogram equalization, binary level slice, piecewise linear
transformation, local adaptive histogram equalization (LAHE) and local
adaptive histogram normalization (Wallis).

Algebraic operations - Add, subtract, multiply, divide, square root and
logarithm of gray scale images.

Geometric operations - Image translation, rotation and scaling. A
registration feature allows interactive alignment of pairs of images.

Filtering operations - Smoothing and median filtering. Kirsch, Laplace,
Roberts, sharpening, sigma and Sobel edge detection operators. Special
line detect filter for enhancing thin fibrous structures.

Morphological operators - Area fill, dilate, erode, boundary track and
medial axis operators for delineating boundaries, identifying objects and
creating overlay masks to be used for region-of-interest operations during
radiometric and filtering operations.

Interactive Graphics Editing

ContourEdit lets you edit graphics data in the form of points, lines and
closed contours. Data can come from a variety of sources including a
digitizing tablet, boundaries of objects extracted from images by the
Image program or the MD2 Microscope Digitizer from Minnesota Datametrics.
You can view your data interactively in 3D from any vantage point using
orthographic or perspective projection. Individual vertices, points, lines
or contours can be selected using hardware picking and then visually
translated, rotated or scaled and saved in their new position. Vertices
can be deleted from or added to line and contour elements. Lines and
contours can be copied or created as interpolations of adjacent lines or
contours. Set operations are one of the most powerful features of
ContourEdit. Elements can be added to or removed from named sets of
elements and each set treated as a single geometric object. Set name
information is preserved in the database and individual sets can be
written to their own disk files.

Isosurface Extraction and 3D Rendering

Cubes analyzes your serial section data, be it digitized gray scale images
or processed contour and point data, and writes a 3D geometry file
containing a winged-edge, linked list of surface polygons and vertex
normal vectors. Voxel values can also be extracted from image data and
saved in the geometry file.

The Display program renders your 3D model using sophisticated graphics
techniques:

Ambient, diffuse, specular, transparency and emissive properties by simply
choosing a material from a supplied library of materials - or design your
own materials.

Multiple, colored light sources.

Phong shading for seamless rendering of surfaces.

Smoothing of surfaces.

Interactive translation, rotation, scaling and spinning of 3D models.

Automated animation of rotation sequences with frame-by-frame screen
capture to disk files.

Volumetric Analysis

The Metrics program allows you to measure the numbers of objects, surface
areas and volumes of all or a portion of a 3D model using a bounding box.
The bounding box can also be used to create cutaway views.

DField computes distance fields of 3D surface models and intersections
between 3D surface models.

Utility Functions

The full-featured version of ImageVolumes is supplied with a number of
utilities for image format conversion, 3D geometry file format conversion
and for capturing screen images to disk in tagged-image-file-format
(TIFF). Image conversions include TIFF, PCX, sample-scanline and PIC.
Geometry file conversions include AutoCAD DXF.

Imagist2
--------
From Princeton Gamma Tech-integrated microscope and analysis systems.
Platforms: Sun. Contact: Princeton Gamma Tech, 1200 State Road, Princeton,
NJ 08540.

IRAF
----
Image Reduction and Analysis Facility. National Optical Astronomy
Observatory (NOAO) Contact: <mailto:[e-mail]>.

IRIS Explorer
-------------
IRIS Explorer was originally developed by Silicon Graphics for their
workstations. It is a modular visualisation environment - you create your
application interactively by connecting modules together using a
point-and-click interface. IRIS Explorer comes with about 150 modules
(more are available) which perform tasks such as reading in data,
filtering it, transforming it; creating graphical objects like line
graphs, histograms, contours, surfaces, isosurfaces, volumes, vector
plots, etc; and displaying them together in a window with full 3D
interaction. A number of modules are built using Silicon Graphics'
ImageVision library, and provide a large amount of image processing
functionality.

You can create your own modules to read or translate data using a
point-and-click tool called the DataScribe, or use the Module Builder -
another tool bundled with the system - to transform your existing routines
(in the form of C, C++ or FORTRAN source, or even as pure executables)
into modules for use from within IRIS Explorer. Finally, IRIS Explorer
provides the application developer with the ability to customise the look
and feel of the application before handing it over to the end-user.

Recently, SGI licenced IRIS Explorer to the Numerical Algorithms Group
(NAG), who are porting it to Sun, IBM RS/6000, HP and DEC platforms. The
Sun and RS/6000 ports are available 10/94; the others will follow soon.
Please contact the IRIS Explorer Centers for more details.

WWW: <http://www.nag.co.uk:70/1h/Welcome_IEC>. Usenet: comp.sys.sgi,
comp.graphics.explorer. <ftp://ftp.epcc.ed.ac.uk/pub/explorer/> or
<ftp://swedishchef.lerc.nasa.gov/explorer/>.

IRIS Explorer Center (Europe): PO Box 50, Oxford OX2 8JU, UK, Tel: +44
(0)1865 516377, Fax: [phone], <mailto:[e-mail]> and
<mailto:[e-mail]>.
IRIS Explorer Center (North America): 1400 Opus Place, Suite 200, Downers
Grove IL 60551-5702, USA, Tel: [phone], Fax: [phone],
<mailto:[e-mail]>.
IRIS Explorer Center Japan (IECJ): Nagashima Building 2F, 2-24-3 Higashi,
Shibuya-ku, Tokyo, Japan, Tel: [phone], Fax: [phone],
<mailto:[e-mail]>.

KBVision
--------
Software environment for creating image understanding applications-
automatic detection, classification, and statisticas generation.
Platforms: Sun, IBM (RS6000s), DEC, SGI. Contact: Amerinex Artificial
Intelligence, 39135 Walnut Terrace, Fremont, CA 94536, Phone:
510-794-7853, Fax: [phone].

Khoros
------
KHOROS version 1.0 does not do 3D visualization, version 2.0, scheduled to
be released 8/94 might have 3D Tools. The small contributed toolbox in
KHOROS v 1.0 does very nice surface thresholding using image gradient
technique and produced grayscale renderings of surfaces. Sun, SGI, DEC,
HP, IBM, NeXT. The three major requirements for Khoros are: X11R4, a
UNIX-type operating system, and lots of space (min. 120-150 Meg). Only if
your PC has these three prerequisites can you consider doing a port of
Khoros to your PC. Successful ports of Khoros have been done for the Mac
II and various 386/486 machines that meet these requirements. Talk to
Donna Koechner <mailto:[e-mail]>. Contact:
[e-mail]. Usenet: comp.soft-sys.khoros.

<ftp://ftp.eece.unm.edu/pub/khoros/>
<ftp://ftp.uu.net/pub/window-sys/khoros/>
<ftp://popeye.genie.uottawa.ca/pub/khoros/>
<ftp://ftp.rrz.Uni-Koeln.DE/graph/khoros/>
<ftp://ftp.lrz-muenchen.de/local/khoros/>
<ftp://ipifidpt.difi.unipi.it/pub/khoros/>
<ftp://ftp.waseda.ac.jp/pub/khoros/>
<ftp://ftp.mcc.ac.uk/pub/cgu/khoros/>
<ftp://unix.hensa.ac.uk/pub/uunet/window-sys/khoros/>
Pull back the file $KHOROS_FTP/release/install.ftp and read it first.

>Alex Milshteyn: The toolbox written by that guy from Italy works fine for
me. Khoros is able to output data in .rs format, therefore Sunview would
handle it just fine as well.

MacCubeView
-----------
Designed to display a texture map image of three-dimensional (3-D) data.
In this release, three simple ray-tracing techniques have been added. The
data in mind is typically generated by medical imaging techniques such as
CT, MRI, and nuclear medicine. Some geophysics techniques also produce
suitable 3-D image data. Hardware Requirements: MacCubeView will probably
run on any Colour Macintosh Computer that is running System 7 or newer.
The software will be at its best when used with a large eight-bit colour
monitor. The Macintosh should have at least eight megabytes of memory
installed. It will not run on a Mac Plus, SE, Classic, etc. Two versions
of the programme are supplied - one for machines with a FPU, the other is
for machines without a FPU, such the LC475 and Quadra 605. There is a demo
file of a slice of the author's head.
<ftp://mac.archive.umich.edu
/mac/graphics/graphicsutil/maccubeview1.50.sit.hqx> and the mirrors around
the world.

MacPhase
--------
2D data analysis and visualization application for the Macintosh. Data
sets can be byte, integer, longint, or real and can be as large as memory
allows. MacPhase has an extensive collection of processing tools ranging
from simple math operators to fourier transforms. You can use simple tools
to filter in the frequency domain. There are 3x3 and 5x5 configurable
convolution filters and much more. MacPhase can also display your data
using raster, contour, 3D wireframe, 3D surface, 3D rendered surface,
vector, 3D contour, 3D line, line, and combination plots. There is an easy
to use color look up table editor. Use it to put the colors where they
best show your data. MacPhase has a Data Tool palette which allows you to
draw in the data layer of your data window. MacPhase has also has a Draw
Tool palette which draws in an drawing layer of your data window. Use the
draw tools to annotate your data with text, simple shapes, placed
pictures, color look up table legends, and even sound objects. Just about
every function or operation can be called using a pascal-like macro
language. Macros are a great way to extend some of the already great
features in MacPhase. You can also write external code modules, Add-ons,
for MacPhase. These Add-ons can be an excellant way to extend MacPhase's
capablities. Add-ons will be made available to support the QuickCapture
and SCION frame grabbers, GPIB interface, serial ports, QuickTime, video
digitization (AV-vdig), Photoshop plug-ins, color channels, image
restoration, and more. Add-ons are callable from the macro language.
Add-ons can be used to add new file formats as well. MacPhase has a large
number of supported formats some of which are PICT, TIFF, MatLab, HDF,
FITS, binary, text, EPS, Mathematica, Photoshop, polygon files, sound,
color tables, and others. MacPhase supports AppleEvents through the
DoScript event and several custom events. Use the DoScript event to send
macro commands to MacPhase. Use the custom events to pass data between
applications.

<ftp://sumex-aim.stanford.edu/info-mac/sci/mac-phase-20-nofpu-demo.hqx>
and /mac-phase-20-demo.hqx (and the mirrors around the world). Contact:
Doug Norton, Otter Solution, 10 Limekiln Road, Whitesboro, NY 13492-2338,
USA, Phone: [phone], Fax: [phone], Internet:
<mailto:[e-mail]>, American Online: OtterSol, AppleLink: ottersol.

MacStereology
-------------
Demo version is available at
<ftp://zippy.nimh.nih.gov/pub/nih-image/programs/>. MacStereology is a
package designed to make measurements of images and to make 3-D
reconstructions. Input to MacStereology is either from a digitising tablet
or from Pict files. The boundaries of the objects of interest can
therefore be drawn by hand on the tablet or traced automatically on a
binary image. From these boundaries and the magnification, parameters such
as area, perimeter and centre of gravity are calculated. If the
co-ordinates of each boundary are also saved, together with the section
thicknesses then 3-D reconstructions can be displayed, printed or plotted,
using a wireframe (for pen plotter), layers or surface plot. MacStereology
should work with any Macintosh with at least 1 Mbyte memory. It was
designed for a MacII with 8-bit colour, but is OK in grey tones or black
and white.

>Stephen M Echteler: [...]MacStereology to do some 3D reconstructions of
developing sensory neurons. The program is rather expensive ($750) and the
Mac interface is a bit buggy. I'd really appreciate comments from anyone
who:1) has used this program or 2) could suggest an alternative
application with similar features.

>John Russ: Well, I've been a MacStereology user for several years now. We
use it in our research (3D reconstructions from all kinds of imaging
including TEM, confocal light, and x-ray microtomography), as well as in
teaching courses to grad students, and like it a lot. There are only three
basic approaches to 3D reconstruction: a) volumetric (transparency)
imaging like VoxelView or VoxBlast, which shows all of the data, but can
be VERY time consuming to fiddle with all of the transparency, lighting,
etc., parameters to reveal the important aspects of structure (they really
require you to already know what is there, and just use the program to
show it to others); b) resectioning approaches like Spyglass Dicer, which
allows you to examine arbitrary sections but cannot show the important
topological characteristics present in the 3D volume; and c) surface
rendering, as in MacStereology, which is very efficient (small files and
fast displays), shows the topology and presents images that appear natural
because we are all used to seeing surfaces, but accomplishes this by
hiding other detail including internal structures behind the surfaces. The
three approaches are complementary and we use them all, but if I had to
choose, I would take Macstereology first, Spyglass Dicer a very close
second, and Voxelview (or Voxblast) a distant third, based on the amount
they are used, and the response of students and researchers to the images
(how much they can learn from them, how difficult it is to interact with
them, etc.).
As to the two specific complaints: I don't agree that the interface is
"buggy." It does have a few peculiarities that are not totally Mac-like,
like fiddling with the display LUT and taking over the whole window, but
you can turn that off if you like. Whenever I've found a bug (usually when
Apple releases a system upgrade or new hardware), the author has fixed it
pretty quickly, and he is also very good about giving advice via e-mail.
And the complaint about the price is really sort of annoying. Photoshop
costs nearly as much, but consider the number of copies they sell? What do
you think Spyglass' set of programs cost? Or how about Voxelview which is
considerably more expensive? How much did you spend for your computer+
camera+ interface+ microscope+ printer+... - well you get the idea. $750
for a program that has taken man-years to develop and has a very
specialized market is hardly high-priced. You are just spoiled because
Image is free (well, unless you count that we all pay taxes to support
Wayne). In the PC world, you would spend $2K or more for a program
equivalent to Image. Expensive? No, expensive is trying to do without a
tool you need.

MCID
----
Image analysis and quantification mainly for fluorescence imaging.
Platform: PC. Contact: Imaging Research Inc, Brock University, 500
Glenridge Ave, St. Catharines, Ontario, Canada L2S-3A1, Phone:
905-688-2040, Fax: [phone].

MetaMorph
---------
Integrated microscope image capture, enhancement, reconstruction, and
visualization system. Platform: PC. Contact: Universal Imaging
Corporation, 502 Brandywine Parkway, West Chester, PA 19380.

MicroVoxel
----------
OS/2 v2.1. Indec Systems, Inc. 820, Bay Avenue #212, Capitola, CA 95010,
[phone].

>Jeff Ingeman: MicroVoxel is a 3D imaging package that imports data from
BioRad MRC-600 files, TIFF files, or raw 8-bit data. You can visually
examine any slice made at any angle or plane through the 3D volume. You
can render your 3D data in 3 different modes of volume rendering. You can
also extract objects from your data and render them in shaded-surface
mode. There are also a number of 2D and 3D image processing tools included
in the program. Animated movies can be rendered and shared with others
using an included, public-domain viewer program. Multiple volumes of 3D
data can also be merged into a tricolor shaded rendering. Markers can be
placed anywhere in 3D space and numerous measurements taken. We have been
using it here at UCI for over a year now and are quite happy with it.

Montage
-------
One of the first complete serial-section reconstruction packages and was
produced at the University of Pennsylvania. It includes component programs
for 2D data entry from digitizer, 3D reconstruction and display, and
surface area/volume analysis. Platforms: PC (Linux w/ VGA or X11), Unix
workstation (Sun, IBM, SGI, etc.). Cost: Free (scientific community), $ if
extended support required.
<ftp://retina.anatomy.upenn.edu/pub/mont.linux.tgz>. Refs: Journal of
Neuroscience Methods v21, pp 55-69, 1987. Contact: Robert G. Smith,
Department of Neuroscience, University of Pennsylvania, Philadelphia, PA
19104-6058, <mailto:[e-mail]>.

Neurolucida
-----------
Low-end interactive image analysis software for neuron tracing and
anatomical mapping. Platforms: PC (Windows). Contact: MicroBrightField,
Inc, 75 Hegeman Ave, Colchester, VT 05446, Phone: [phone], Fax:
802-655-4031.

NCSA Tool Suite
---------------
Unix Workstations (DEC, IBM, SGI, Sun), Macintosh, Cray. Contact: National
Center for Supercomputing Applications, Computing Applications Building,
605 E. Springfield Ave., Champaign, IL 61820. Cost: Free. The suite
includes tools for 2D image and 3D scene analysis and visualization. The
code is actively maintained and updated. <ftp://ftp.ncsa.uiuc.edu/>.

NCSA Data Slice (XDataSlice). Supports X11. /UNIX/XDataSlice.

Viewit in /misc/viewit/. Viewit is a memory hog which can do array
manipulations on entire 3d datasets, some limited format conversions, and
3D volumetric projections. This program includes facilities for
constructing animated sequences of 3D volumetric views through various
volumes. Viewit has been ported to a number of machines including Crays
under UNICOS, Sun Workstations, Silicon Graphics Workstations, Alliant and
Convex Minisupercomputers, and a variety of other machines. Public domain.
Contact: [e-mail].

Tiller in /misc/tiller/. Tiller is an SGI viewer for viewit displays.

>Patrick Moran: Viewit has a command line interface and is not the
friendliest software that you can use. On the other hand it does have a
large number of built-in functions that are useful for image processing,
including functions for MR reconstruction. Viewit is used by researchers
doing work in MR here.

>Joe Biegel: I never said Viewit had a great User I/F - in fact it
doesn't! However, it does MANY things including volume rendering quite
well. If you read the documentation, it's not that hard to get results.
It's also free. I've been using it for a few years - it's not NIH Image,
but it DOES do things like depth cued volume rendering. I've used it quite
a bit for brain imaging visualization - it works - there is a learning
curve, but it works. If you want more info on ViewIt, send email to
[e-mail] (Clint Potter).

>Alexander-James Annala: Check out the code for NCSA viewit (if you have
an unix/x11 platform) -- requires 'tons' of memory for any reasonable size
volume -- and it is not exactly GUI based code (uses Tk command line
tools) -- but it does have a huge manual, 'tons' of functionality, and it
is fast when run on systems with fast processors and 'tons' of memory.

>Alexander-James Annala: If you have a SUN SPARC workstation/server with
>250M free memory available (that's free memory -- not available disk
space) then you can use NCSA's Viewit (NMR Imaging and Spectroscopy
Package) to do 3d volumetric imaging at full resolution of the UNC CHVRTD
datasets.
The following SUN SPARC recipe displays multiple views of a 3D head: get
any X11R5 (or maybe X11R4) server running on myhost - this is where you
are going to display images -- rlogin to bighost - this is where you will
need the free memory for storing intermediate results during the volume
rendering.
myhost(96): rlogin bighost -l myusername
bighost(1): ftp -i ftp.ncsa.uiuc.edu
ftp> cd misc/viewit/viewit.v3.13
ftp> binary
ftp> get viewit.sparc-version.Z
ftp> get viewit.help
ftp> quit
bighost(2): zcat viewit.sparc-version.Z >viewit
bighost(3): chmod 755 viewit
bighost(4): ftp -i omicron.cs.unc.edu
ftp> cd pub/softlab/CHVRTD/volI
ftp> binary
ftp> get 3dhead
ftp> quit
bighost(5): dd if=3dhead of=3dhead.s conv=swab
bighost(6): viewit
viewit(tcl)> -dim 3 256 256 109 -iformat RAW -itype USHORT -i 3dhead.s
viewit(tcl)> -scale 1.0 1.0 2.3486 -reorder 2 0 1 -push
viewit(tcl)> -dim 1 90 -ramp -1 1 -linscl 0 356 -xchg
viewit(tcl)> -movie 0.0 add no_erode
viewit(tcl)> -linscl 0 255
viewit(tcl)> -displ X myhost
bighost(7): exit
myhost(97): logout
If you have a smaller workstation you can still do some limited volume
rendering - but you will have to subsample the original 3D dataset to
reduce swapping to disk to a reasonable level.
A viewit electronic newsletter is distributed on an irregular basis -
email to <mailto:[e-mail]> to request a subscription. Clint
Potter (the original author and pricipal contact) is at
<mailto:[e-mail]>.

NIH-Image
---------
Has painting and image manipulation tools, a macro language, tools for
measuring areas, distances and angles, and for counting things. Using a
frame grabber card, it can record sequences of images to be played back as
a movie. It can invoke user-defined convolution matrix filters, such as
Gaussian. It can import raw data in tab-delimited ASCII, or as 1 or 2-byte
quantities. It also does histograms and even 3-D plots. It is limited to
8-bits/pixel, though the 8 bits map into a color lookup table. It runs on
any Mac that has a 256-color screen. Free.

DICOM import routine. In addition to DICOM-3 images, it now reads many
ACR/NEMA images. It now requires a DICOM dictionary to decode the DICOM or
ACR/NEMA header. The dictionary is available from
<ftp://zippy.nimh.nih.gov/pub/nih-image/documents/>. Hold the option key
down to get a full dump of the DICOM header.

<ftp://zippy.nimh.nih.gov/pub/nih-image/>. Mailing list:
<mailto:[e-mail]>, subscriptions to the mailing list:
<mailto:[e-mail]> (subscribe nih-image "your name") (set
nih-image mail digest). The best way to search the archived messages is to
use <gopher://saturn.soils.umn.edu:151/77/wais-sources/nih-image-archive>
to search the archived messages for keywords, or
<gopher://saturn.soils.umn.edu/11/email-lists/nih-image> for the actual
archived messages. The archived messages are also available at
<ftp://ftp.soils.umn.edu/pub/info/email-lists/nih-image/>. The archived
messages are also in <ftp://zippy.nimh.nih.gov/pub/nih-image/documents/>
but some of the more recent ones may be missing. You can also obtain a
list of the available archive files by sending an "index nih-image"
command to <mailto:[e-mail]>. These files can then be
retrieved by means of a "get nih-image filename" command.

Nuages
------
Input: a set of simple closed polygons on parallel planes. There may be
several (nested) polygons per plane. Output: A set of triangles
representing the surface of a 3D polyhedra, and/or a set of tetrahedra
filling the 3D polyhedra. The program adds vertices onto and inside the
contours. How to display the output: The program currently supports
wavefront .obj format, DXF format and Object File Format (.off) format.
The latter can be visualized with geomview on sgi and NeXT. Geomview is
available at geom.umn.edu. Conversion tools to other file formats and
several contour sets can be found at <ftp://avalon.chinalake.navy.mil/>.
How to get input data: The input format is a simple ascii file (see man
prepros for a format description).

Platforms: Sun, SGI, DEC. Refs: "Three-dimensional modeling of human
organs and its application to diagnosis and surgical planning.", Technical
Report 2105, Institut National de Recherche en Informatique et
Automatique, (France), Dec 1993. <ftp://betelgeuse.inria.fr/pub/Nuages/>
(138.96.16.91).

NUAGES_SUN4.tar.Z     sun sparc 2 (sun4OS4)
NUAGES_SGI.tar.Z      sgi         (iris4d)
NUAGES_DEC.tar.Z      decstation  (ultrix)

>Matthew T. Adams: I wanted to let you know about a little utility I just
wrote and posted to <ftp://ftp.uwa.edu.au/pub/povray/incoming/utilities/>.
It converts Bernhard Geiger's Nuages' "vera" file format into the PoV 2.x
file format, suitable for inclusion into a PoV scene file.

OLPARS
------
On-Line Pattern Analysis and Recognition System from the PAR Government
Systems Corporation. Statistical pattern recognition system that can be
applied to analyzing information derived from reconstructions. Platforms:
Sun (Unix), DEC (VAX/VMS). Contact: Amber Technologies, 47 Junction Square
Drive, Concord, MA 01742, Phone: [phone], Fax: [phone].

OSIRIS
------
OSIRIS has been designed as a general medical image manipulation and
analysis software. The design is mainly based on the following criteria:
portability, extendibility and suitability for any imaging modality.
OSIRIS is designed to deal with images provided by anytype of digital
imaging modality to allow physicians to easily display and manipulate
images from different imaging sources using a single generic software
program. Portability ensures the software implementation on different
types of computers and workstations. Thus, the user can work in the same
way, with exactly the same graphical user interface, on different
stations. Also by supporting standard file formats, the OSIRIS software
provides access to images from any imaging modality. The OSIRIS program
was developed as part of the Geneva PACS project and is intended for
physicians and non computer-oriented users allowing them to display and
manipulate medical images. Its standard original version included only
basic image manipulation tools accessible through a convenient and
user-friendly graphic interface. In addition to being used at the
University Hospital of Geneva, it was widely distributed around the world
and was adjusted according to user's comments and suggestions. This
program was also designed to serve as a development of more advanced image
processing and analysis tools.

Portability: The initial development of OSIRIS program was undertaken
simultaneously on UNIX based X/window graphic environment as well as Apple
Macintosh native platform. The UNIX version was further tested on a
variety of workstations namely SUN Sparc series, DEC alpha series, HP 7000
series, IBM Risc-6000 series and SGI machines. Recent evolution in the
desktop computing environment lead us to develop a new kernel of the
OSIRIS software to be compatible with Windows 4 graphic interface for PC
compatible computers. This new version is also directly compatible with
Windows NT and may be used on Windows 3.1 with some restrictions in
performance. Finally we also ported our code to run native on the PowerPC
RISC computers to fully benefit from the enhanced performance of this new
generation of processors.

OSIRIS provides (just a few characteristics): Interactive graphic user
interface, Customizable display modes for images sets, Zoom, rotation,
flipping of image sets, Color adjustment on full dynamic range, Magnifying
glass, Annotations, Regions of interest (polygons, ...), Measurements
(distance, angle, surface, volume, ...), Filters, Multiplanar sections of
tomographic images, Region growing for automatic image segmentation,
Histogram equalization.

OSIRIS software can be obtained free of charge from: Digital Imaging Unit,
University Hospital of Geneva, 24 Micheli du Crest, 1211 Geneva 14 -
Switzerland. Fax: [phone], <mailto:[e-mail]>. A
special developer license is available for the full source code.
<http://expasy.hcuge.ch/www/UIN/osiris.html>,
<ftp://expasy.hcuge.ch/pub/Osiris/>.

Pixar
-----
High-end visualization and rendering for movies, but also for the medical
community. Contact: Pixar, 3240 Kerner Blvd, San Rafael, CA 94901, Phone:
415-258-8100, Fax: [phone].

Pixcell
-------
$1,500 from Sandia Labs. A demo version of Pixcell, complete with manuals
and images from Sandia Labs <ftp://ecto.ca.sandia.gov/pub/Pixcell/>.

PV-WAVE
-------
PV-WAVE is a comprehensive software environment that integrates
state-of-the-art graphical and numerical analysis techniques into an easy
to use, easy to extend, easy to apply, and easy to learn system for
quickly finding solutions to, and building applications for, complex
technical problems.

PV-WAVE uses an intuitive fourth-generation language (4GL) that analyzes
and displays data as you enter commands.  With it you can perform complex
analysis, visualization, and application development quickly and
interactively.

PV-WAVE provides hundreds of routines for representing, importing,
exporting, filtering, transforming, analyzing, visualizing, and
communicating data, as well as constructing widget-based applications with
this technology.

PV-WAVE is available for UNIX and OpenVMS workstations and for PCs running
MicroSoft Windows or Windows NT.

Cost: $995 to $6995.  Education and quantity discounts available.

Contact: Visual Numerics, Inc., 6230 Lookout Rd, Boulder, CO 80301, Phone:
800-447-7147 (outside the US: 303-530-9000), Fax: [phone].

<mailto:[e-mail]>, <http://doc:8118/vnihome.html>, Usenet:
comp.lang.idl-pvwave.

RMN
---
>Philip Grandinetti <mailto:[e-mail]>: RMN - A Nuclear Magnetic
Resonance (NMR) data processing program for the Macintosh. Free.
<ftp://ftp.funet.fi/pub/sci/chem/nmr/>.

Various types of one- and two-dimensional NMR data processing. It can read
in text files (single column of data), and data files from most
Chemagnetics and Varian spectrometers. It should also read in Tecmag data
files. As far as Bruker data files are concerned, I could never figure out
what their format is. If anyone has specific details, please let me know.

There are two versions: "RMN" - for Macs with a math coprocessor, and "RMN
(no 881)" - for those without. "RMN (no 881)" will run on a PowerPC,
however, it is certainly not native mode.

There is no manual for this program. Most menu items should be obvious.
Just try it and see what happens. The menu Analyze is disabled until I
have time to fix some bugs. The same is true for Simulate under the
Acquire menu.

ROSS
----
Reconstruction Of Serial Sections. Serial-section based reconstruction and
visualization system for microscopy. Interactive and automated mosaicking,
contour extraction, and registration. Platforms: SGI. Cost: free.
Availability: late `94. Contact: Dr. Muriel Ross, Biocomputation Center,
MS239-11, NASA Ames Research Center, Moffett Field, CA 94035-1000,
<mailto:[e-mail]>.

SciAn
-----
SGI 4D, IBM (req. GL, Z-buff). <ftp://ftp.scri.fsu.edu/pub/SciAn/> or
<ftp://monu1.cc.monash.edu.au/pub/SciAn/>.

>Bob Lipman: [where to get red-blue 3D glasses] So far I've heard about 3
places: Reel-3D, Culver City, CA 310-837-2368, American Paper Optics,
800-767-8427 and Cygnus Graphic, Phoenix, AZ. Some of you asked about the
software that displayed the red-blue image. The software is called SciAn.

Semper6
-------
General image Processing and acquisition system. Platforms: PC, DEC (VAX),
Sun. Contact: Synoptics Ltd, Paragon Towers, 233 Needham St, Newton, MA
02164, Phone: [phone], Fax: [phone].

SNARK93
-------
SNARK93 - a programming system for 2-D image reconstruction from
projections for the UNIX/Sun environment, is available from the Medical
image processing group, Dept. of radiology - University of Pennsylvania.

SNARK93 is a programming system designed to help researchers interested in
developing and evaluating reconstruction algorithms for image
reconstruction from projections. It is the latest in a series of releases
of SNARK. One of these, SNARK77, is described in some detail in the book
by G.T. Herman, "Image Reconstruction from Projections: The Fundamentals
of Computerized Tomography," Academic Press, New York, 1980. In fact, all
illustrations of two dimensional reconstructions (by a large variety of
algorithms) in that book were produced by SNARK77. Additional
reconstruction algorithms can be found in SNARK93, such as the linogram
method of Edholm, Herman, and Roberts (IEEE Trans. on Med. Imaging, vol.
7, pp. 239-246, 1987), the maximum likelihood EM algorithm of Shepp and
Vardi (IEEE Trans. Med. Imaging, vol. 6, pp. 113-122, 1982), and the
maximum a posteriori probability algorithm of Herman, De Pierro, Gai (J.
Visual Comm. and Image Proc., vol. 3, pp. 316-324, 1992). SNARK93 also
provides a methodology for testing for statistically significant
task-specific performance differences between algorithms, as illustrated
in the papers by Herman and Odhner (IEEE Trans. Med. Imaging, vol. 10, pp.
336-346, 1991) and Herman and Meyer (IEEE Trans. Med. Imaging, vol. 12,
pp. 600-609,1992). It also extends the capability of previous SNARK
releases (which simulate data collection in X-ray computed tomography) to
emission tomography. SNARK93 has been designed to be flexible and
transportable, in places at the expense efficiency. While it may also be
used to reconstruct repeatedly from data collected by a particular device,
a special purpose program for that device is likely to be much more
efficient.

The SNARK93 programming system is designed to:

(a) be capable of generating mathematically described phantoms
realistically representing various cross-sections of the human body;

(b) be capable of generating mathematically simulated projection data of
such cross-sections reflecting the characteristics (geometrical
arrangements of sources and detectors, spectra, noise properties, etc.) of
various possible tomographic data-collection devices;

(c) contain many of the published reconstruction algorithms;

(d) contain subroutines to carry out work which appears to be common to
many reconstruction algorithms, so as to facilitate the incorporation of
additional (user-defined) algorithms;

(e) contain routines for the evaluation of single reconstructions and
provide a methodology for testing for statistically significant
differences between reconstruction algorithms;

(f) be capable of displaying the reconstructed images and plotting several
distance measures between the original object and the reconstructed image.

SNARK93 will be make available to all who request it at the cost of
reproduction and mailing of the FORTRAN source code (on a UNIX tar tape)
and the manual. The software and the manual may also be received via ftp
(in which case we will require a login ID and a password). We charge
US$200 (checks only; drawn on a U.S. bank) for providing this service. For
overseas mailing add another US$50.00 if air mail delivery is required.
Please make the check payable to RADIOLOGY ASSOCIATES and send it with
your order to:

Ms. Mary Blue, Medical Image Processing Group, Dept. of Radiology - Univ.
of Pennsylvania, Blockley Hall, 4th Floor, 418 Service Drive,
Philadelphia, PA 19104-6021, U.S.A., Tel. [phone], Fax (215)
898-9145, <mailto:[e-mail]>.

SNARK93 CONFIDENTIALITY AGREEMENT & DISCLAIMER

By purchasing the SNARK93 software, the recipient agrees to abide by the
following terms:

1. SNARK93 shall not be redistributed in any way.
2. SNARK93 is not a patient-care tool and it is not approved by the United
States Federal Drug Administration.
3. While every effort has been made to correct all known bugs, SNARK93 is
provided "as is" with no warranty whatsoever. As such, the recipient
agrees not to hold the authors responsible for any problems they may
encounter with the software.
4. The recipient agrees to purchase SNARK93 with the explicit knowledge
that the authors do not offer technical support.

Spyglass Slicer/Dicer
---------------------
Volumetric visual data analysis package that allows users to create 3D
color images in seconds. It can display isosurfaces, cut through data with
oblique slices, and scale, annotate and label the display of data. Slicer
for Windows also has contains translucency and lighting options.

Availability: Macintosh, Windows. List: $495 for Windows, $695 for
Macintosh. Requirements: Macintosh: 256 color display, 4MB RAM; Windows:
256 color display, 8MB RAM, FPU; UNIX: 256 color display, 16MB RAM.

Contact: Spyglass, Inc, P.O. Box 6388, Champaign, IL 61826, USA (or 1800
Woodfield Drive, Savoy, IL 61874?), Phone: [phone], FAX: (217)
355-8925. Spyglass Technical Support at <mailto:[e-mail]>.
<http://www.spyglass.com/>.

>James Sneyd: I've used Spyglass for many years, and found it to be
excellent. Especially when combined with the HDF format libraries.
Animation is excellent, and Mac-Unix coordination is seamless. It's not
nearly as powerful as some visualisation systems (Explorer, AVS,
Wavefront) but is a LOT cheaper and I have never needed anything more
expensive. I use it mostly for numerical solutions of PDEs in two and
three dimensions. One problem is the absence of Dicer for Unix.

Sunview
-------
From SUN. <http://www.sun.com>. Contact: Sun Microsystems, 2550 Garcia
Avenue, Mountain View, CA 94043, Phone: [phone].

SunVision
---------
Sun (SunOS under X). <http://www.sun.com>. Contact: Sun Microsystems, 2550
Garcia Avenue, Mountain View, CA 94043, Phone: [phone].

>Matthew T. Adams: I have a SparcStation 2 (SunOS 4.1.3) with OpenWindows
3, and I use SunVision 1.1 to do MRI volume renderings.
Advantages: easily customizable (interactive GUI editor), you can hook in
your own C code to the GUIs, volume renderings (SunVoxel), lots of image
processing tools (SunIP), photorealistic rendering (SunART, using Pixar's
RenderMan), geometric renderings (SunGV), animation (SunMovie), C library
containing all tools in all the above modules, straightforward file format
(for volume & image data, at least). Drawbacks: SunVision 1.1 is the last
version -no new stuff. Sun recommends SunVideo, speed (I'm not sure if
it's slow because of sloppy coding or my slow machine): ~3 minutes to
render a 256x256x50 8-bit volume, ~12-15 minutes to render a 256x256x124 8
bit volume.

Synu
----
SGI. Contact: [e-mail]. Non-interactive components run on some
Unix-based systems.

>Harvey Karten: There is an excellent new program, called SYNU, that does
elegant 3D reconstruction of neurons, written by the Imaging Group at the
University of California Sand Diego, under the direction of Mark Ellisman.
It runs on a Silicon Graphics machine, and produces gorgeous images of
serial sections, with variable transparency, stereo pairs, etc. I think it
may be available for just the cost of the media. An example of the product
is shown on the front cover of the November issue of J. Neurosciences by
Martone. The current problem with it (when I last spoke with the Ellisman
group about this) was that it takes a bit of doing to import files into it
from Image or Canvas or other programs, and it does need a Silicon
Graphics to run the program.

The Explorer
------------
Macintosh. UCLA.

Theraview
---------

View
----
SGI (4D, Indigo, Crimson). <ftp://ftp.cs.unc.edu/pub/VIEW/>. Contact:
<mailto:[e-mail]> or <mailto:[e-mail]>.

VIS-5D
------
SGI, IBM, SUN, HP, DEC (volume rend. only on SGI).
<ftp://iris.ssec.wisc.edu/pub/vis5d/>
<http://ssec.wisc.edu/~billh/vis.html>
Contact: <mailto:[e-mail]> or <mailto:[e-mail]>.

VIS-5D is primarily designed for interactive visualization of time-varying
multi-variate 3-D gridded data such as the output of numerical simulations
of the atmosphere and oceans.  It can be applied to a variety of other 3-D
gridded data.

Vis-AD
------
SGI only.
<ftp://iris.ssec.wisc.edu/pub/visad/>
<http://ssec.wisc.edu/~billh/vis.html>
Contact: <mailto:[e-mail]> or <mailto:[e-mail]>.

VIS-AD is designed for interactively steering and visualizing scientific
computations.  The system includes a high-level interpreted programming
language with links to C and Fortran. Users define data types appropriate
for their applications. The system includes a novel and flexible way for
users to control how their data are displayed.

VolVis
------
<ftp://sbcs.sunysb.edu/pub/volvis/>. The major restriction is that you
have to have some sort of a Silicon Graphics workstation in order to run
it. If you do not have and SGI, there is a starbase version at that same
site that will run only on HP's with starbase. VolVis can be compiled on
SPARC and runs fine under SGI, HP (w/Starbase), Sun SPARC (X11R5).
Contact: <mailto:[e-mail]>.

>Dave Wyble: While I'm sure VolVis is a very impressive package, it does
not appear to run on Sun OpenWindows.

Vox-L
-----
MS Windows NT (Intel, Mips, DEC Alpha), MS Windows 3.1, Unix/Motif.

Contact: <mailto:[e-mail]>. There is an extensive archive at
<ftp://ftp.near.net/member/dataspace/>. The demo subdirectory has several
demo disks of the Vox-L Visualizer. The Vox-L Visualizer demo can only
operate on an included 128^3 file, but the actual application is quite
comfortable with 256x256x128 sized volume data of MR and CT scans (see GIF
format images in the images subdirectory). The software is also available
with drivers for Stereographics' CrystalEyes. For a complete visualization
environment, check out the Vox-L Workstation line which can provide a
150mhz Alpha AXP workstation tuned for volume rendering which start under
19,000.

Voxblast
--------
Voxel-based 3d volume rendering system. Macintosh, PC. They have a demo
available via FTP. Contact: Vaytek, Inc., 305 West Lowe, PO Box 732,
Fairfield, Iowa 52556, Phone: [phone], Fax: [phone].

>Alex Colburn: PC options may be slower than other hardware options, but
our version of Voxblast running under Windows, will render at about
800,000 voxels per second on a 486 DX2, w/ 16M RAM. Granted it takes about
3 seconds to render the standard hogheart volume ( 202x132x144) and larger
volumes take longer.

VOXEL-MAN 3D
------------

>Karl Heinz H÷hne: Some weeks ago we had announced the availability of the
VOXEL-MAN 3D interactive atlas of skull and brain. Those who like to have
an impression of its functionality may get interactively generated sample
images via ftp at (134.100.96.5) <ftp://fokus.uke.uni-hamburg.de/>
[anonymous.voxelman.images]. All images are in CompuServe ".gif" format.
Be sure to use BINARY transfer mode! File sizes are in "blocks" of 512
bytes. A sample ftp session can be found in the file
ftp-sample-session.log in the anonymous home directory. The server is a
VAX/VMS system, so some aspects are somewhat special:

common UNIX FTP server     this FTP server
__________________________________________
cd voxelman                cd [.voxelman]
cd voxelman/images         cd [.voxelman.images]
cd ..                      cd [-]

Voxel View
----------
Vital Images (505 N. 4th St., Fairfield, Iowa 52556, tel: [phone],
fax: [phone]) for 3D reconstruction of images. SGI, Mac. User service
at <mailto:[e-mail]>. Email: <mailto:[e-mail]>.
Refs: "Voxels: Data in 3D", Byte, May 1992 issue, pp177-182.

>Mustafa Khokha: Vital Images writes its software so that it goes to the
graphics hardware in the machine to do its voxel renderings. On machines
like the SGI, this is great since SGI machines have fast graphics engines
for just this sort of thing. The renderings are fast. I mean really fast.
I was really amazed the first time that I saw it. This is a key part of
both of these programs. If you have to wait many minutes for a rendering
to occur, then a lot of the manipulations like opacity, lighting, and such
become a lot (I mean a LOT) less useful. You learn so much more by being
able to change these parameters on the fly. If you have to wait, forget
it. I just don't have the patience to play with it, and I think you really
lose a lot of "feeling" for what is going on in your datset. Therefore
make sure that you get to try out a datset that is representative of the
ones that you would be doing in terms of size and complexity. Then see if
you can manipulate the settings (like lower the rendering quality) so that
it renders before your eyes. Then when you see something interesting, up
the quality and see if it really pans out. I know VoxBlast allows you to
do this but I am not sure about VoxelView. This is because VoxBlast goes
through the operating system when it makes its calls so they can "cheat"
and render every other voxel or so. VoxelView doesn't because of its
dependence on the hardware. On the other hand perhaps this is only true
for the Unix versions? Anyway something to be aware of.
Also one problem that we have with VoxelView is that you cannot have
fractional spacing in the Z direction. In other words interpolations are
purely whole numbers in the Z direction. This is a real drag since most of
my datasets do not have aspect ratios that are whole number multiples in
the z compared with x and y. Things may not look right then. Again this is
because VoxelView depends on the graphics architecture and VoxBlast does
not.
Personally, for the short time that I tried VoxBlast, I thought it was
quite nice. Everything is menu driven although you can if you want write
your own scripts. Ease of use is not a problem. I would still compare with
the Mac version that Vital Images makes though so be sure to get your
hands on the demos.

>Michael Cammer: VoxelView does do up to 16 bit renderings where bits can
be grouped. We have tried the new color merging function for double
immunofluorescence which allows for each dataset to be full eight bits.

>Dietmar Reiter: I've got some minor experience on VoxelView Ultra on SGI,
as well as on the Macintosh Version of VoxelView (which I know better,
since our lab has got one). Three basic points are remarkable: (1) Speed
(and therefore interactivity) is minimal on the Macintosh release. It
should be somewhat of an indication, what volume rendering SHOULD be on a
graphics computer as the SGI. Speed lacks tremendously, what makes
interacitve working with datasets (volumes) larger than 1 million voxels
(a cube of 100x100x100 voxels) nearly ipossible. It can be helpful to
threshold out most of the voxels, in other words to keep the rendering
task as minor as possible. (2) Nearly no parameters of the VoxelView Mac
can be given by numbers, but by fancy scroll bars. This inhibits for
example an exact setting of the opacity for discret voxel values. A
manually cursor-drawn curve is the approximation. (3) According to sources
at Vital Images, the Macintosh version will not be developed further, what
is understandable, because of hardware limitations of the 680x0 and the
Mac PC system as a whole. This may probaly change with the upcoming of the
PowerPC. William VanZandt at Vital Images will give give more information.

>Greg Gillen: It appears that NIH Image will open Voxel View PICS
animation files and run them at a much faster rate than Voxel View. The
only problem seems to be the LUT is messed up.

>John Russ: [how you might be annotating your animations and if anyone has
transferered animations to video for presentation at meetings] Yes - I use
the mac version, which is a bit slow but OK on a Quadra. I save the
animations as PICS files which I then convert to MooV format using
ConvertToMovie. This lets me play them as Quicktime movies. I play them
directly from a powerbook using an LCD overhead panel, not through
videotape.

Voxtool
-------
From General Electric. Runs on a Sun Sparcstation with Advantage Windows.

>Matti Haveri <mailto:[e-mail]>: We use Voxtool 1.0.4 in a
Sparcstation (64Mt of RAM) connected via Ethernet to General Electric's
High Speed Advantage and GE Sytec CT-scanners.
Binary segmentation is done through lower and higher HU-thresholds. Shaded
surface display, MIP, RaySum, Integral and Multiplanar reformatting (MPR)
as well as movie loop (if memory permits full 360 degree rotation with 6
degree intervals is sometimes possible). Select/remove object, filter
floaters (user-definable size), scalpel (user-definable cut depth),
erosion, dilation, open bridges and close gaps-functions. Intersection,
union, difference and delta-functions between manipulated models.
Crashes occasionally the whole workstation. Restarting the program or
rebooting clears occasional memory-problems. Can save only two models at a
time and the program sometimes loses link to the original slices which is
very annoying. 3D-images can be saved as individual still-images to disk
in GE's image-format (hope we will be able to open and view this format in
the other platforms as well). In the current version area-measurements
require at least two contiguos slices which is also annoying. If the
3D-model has to use more than 100 512x512 image's space as virtual memory
the program is too slow to use interactively. The current version doesn't
support colors, transparency or simulated surgical interventions. There
are some demo-pictures at
<http://biocomp.arc.nasa.gov/3dreconstruction/movies>.
Contact: We get tech support and upgrades through local GE staff here but
when we have some questions they often consult GE anyway and it takes
time. I have heard that in the past it was possible to contact GE's
software-staff directly but now GE wants that the local staff must be
consulted first. ...wish we had tech support through the Internet, it
would be _much_ easier. GE - are you listening?

VROOM
-----
Contact: Karel Zuiderveld, Computer Vision Research Group, Utrecht
University Hospital, E02.222, Heidelberglaan 100, 3584 CX Utrecht, Phone:
[phone], <mailto:[e-mail]>.

VROOM is an acronym for Volume Rendering using Object-Oriented Methods; it
is a C++ class library aimed at multi-modal visualization. A description
of the functionality can be found in: Zuiderveld KJ, Viergever MA.
Multi-modal Volume Visualization using Object-Oriented  Methods.
Proceedings of the 1994 Symposium on Volume Visualization (Washington
D.C.), ACM SIGGRAPH 1994:59-66.

The VROOM class library runs on various platforms (including HP, IBM
RS/6000, Sun, SGI and Intel-based PC's running UnixWare or Linux) with a
wide variety of compilers (as cfront, g++, xlC, Symantec).

Unfortunately, we can not give VROOM source code to interested parties
(due to restrictions placed upon us by the sponsoring companies). We are
currently working on a Tcl/Tk binding for VROOM; this will allow us to
provide others with a major part of VROOM's functionality since we then
can place TclVroom binaries on our ftp server.

For more information on the Computer Vision Research Group, see our WWW
page.  <http://www.cv.ruu.nl>.

Wavefront Data Visualizer
-------------------------
SGI, SUN, IBM, HP, DEC. Contact: <mailto:[e-mail]>.

WHIP
----
General purpose image processing software from GW Hannaway & Associates.
Also has automated stage control and acquisition capabilities. Platforms:
SGI. Contact: GW Hannaway & Associates, 839 Pearl Street, Boulder, CO
80302, Phone: [phone], Fax: [phone].

XEVA-VisualStudio
-----------------
XEVA-VisualStudio is an interactive system running on most Unix
workstations, supporting visualization of volume data sets from image
slices (acquired by CAT, NMR, Ultrasound, microscopy, confocal laser
systems). The program runs on SGI IRIX 4.5.1, SunOs 4.1.3_U1, HP HPUX
A.09.01, IBM Risc6000 AIX 3.2.5 (shared X11 libraries only).

XEVA-VisualStudio volume rendering system is available for free at
<http://www.dsi.unimi.it/Users/imaging/eva.html>

Zmode
-----
>John Noel [Software/hardware that can convert a series of parallel MRI/CT
scan images to a 3D reconstructive model in a CAD system?]: Zmode has
developed the ability to manufacture models representing patient anatomy.
The models are produced using medical imaging software and rapid
prototyping technology. Zmode provides this modeling as a service and can
also provide custom software solutions such as translation into CAD
entities. You can get more information by mailing
<mailto:[e-mail]>.

Some medical sites
==================

Austin.au PET-Digital Image Library
-----------------------------------
<gopher://gopher.austin.unimelb.edu.au/11/images>. Austin Hospital,
Australia: 3-D PET-images of a human brain etc.

Avalon 3D object files
----------------------
<ftp://avalon.chinalake.navy.mil/pub/objects/>. 3D object files
subdirectories exist for different formats. Avalon was created to be a 3D
object "repository" for the net. You'll find 3D datasets in various
formats, utilities to convert between the different formats, and documents
explaining the file formats. There is a /pub/incoming directory for
uploads, so if you have anything to contribute, please upload it! If you
have any problems connecting to avalon, try its mirror site Kubota Pacific
at (144.52.120.9) <ftp://ftp.kpc.com/pub/mirror/avalon/>.

AVS and IDL in medical treatment planning
-----------------------------------------
<http://archive.xrt.upenn.edu/0h/buhle/manuscripts/avs94_paper.html>.
Hypertext version of paper on using AVS and IDL in medical treatment, by
E. Loren Buhle, Jr. Ph.D.

Chapel Hill Volume/Ray Tracing Datasets
---------------------------------------
<ftp://omicron.cs.unc.edu/pub/softlab/CHVRTD/>. Chapel Hill Volume/Ray
Tracing Dataset. MRI scan of a human skull (256x256x109) and a CT scan of
a skull of a human cadaver (256x256x113). Note that these are not images
but volumes, but the slices could be treated as images if you can extract
them.

Collaborative Hypertext of Radiology and Ultrasonography
--------------------------------------------------------
<http://chorus.rad.mcw.edu/chorus.html>. CHORUS (Collaborative Hypertext
of Radiology and Ultrasonography) is a multi-author, multi-institution
hypertext that contains more than 1100 documents on radiologic anatomy,
differential diagnoses ("gamuts"), and related diseases and syndromes.
CHORUS will incorporate facilities for distributed authoring, peer review,
and publication via the World Wide Web. It's based on Fact/File, a
radiology hypertext that has been integrated with a clinical radiology
information system since 1990. Comments or questions about CHORUS are
welcome at <mailto:[e-mail]>.

Comparison of Visualization Techniques and Packages
---------------------------------------------------
<http://www.sara.nl/Consumer.Report/Report.html>. We have made a
comparison of visualization techniques and packages. It consists of a
general part with descriptions of AVS, IRIS Explorer, Data Explorer (DX)
and Data Visualizer. The other part contains a study about how these
visualization packages have been used in visualizing plasmaphysical
processes.

CRS4
----
Center for Advanced Studies, Research and Development in Sardinia.
Interactive Volume Visualization CT/MRI Data Tools at CRS4, Medical
Imaging Movies at CRS4, Semi-Trasparent Direct Volume Rendering at CRS4,
Fuzzy Gradient Shading Methods at Niguarda Hospital, Simple 2D
Visualization Methods at Niguarda Hospital.

<http://www.crs4.it/~france/vvis.html>.

Blood flow in intracranial arteries and aneurysms
-------------------------------------------------
<http://www.neuronet.pitt.edu:8910/~georgef> or
<http://www.pitt.edu/~gnfst1>

Finite Element Modeling of Intracranial Arterial Blood Flow and Saccular
Aneurysm Formation. Efficient Segmentation and 3-D Reconstruction of
Intracranial Vascular Structures from Magnetic Resonance Angiography Data.

>George Foutrakis (University of Pittsburgh, Department of Neurological
Surgery): I've put together a WWW Home Page describing my past and current
research. The main focus is on the computational modeling of blood flow in
intracranial arteries and aneurysms.

General Electric
----------------
<http://www.ge.com/>
<http://www.ge.com/crd/ivl/three_dim_medical.html>

For the purpose of example image sets, GE images are provided in two
formats: DICOM Part 10 format and Central Test Node (CTN) v2.2
(Mallinckrodt Institute of Radiology) format. The GE DICOM images are
available solely for the purpose of providing examples and do not
represent a complete set of every image type. These images are not meant
for exhaustive testing purposes.

<ftp://ftp.med.ge.com/pub/DICOM/IMAGES/CT/>
<ftp://ftp.med.ge.com/pub/DICOM/IMAGES/CT/HiLgtHsp/PART10/>

DICOM conformance statements:
<ftp://ftp.med.ge.com/pub/DICOM/CONFSTMT/>

University Hospital of Geneva - Digital Imaging Unit
----------------------------------------------------
<http://expasy.hcuge.ch/www/UIN/UIN.html>. UnitO d'Imagerie NumOrique.
Provides new perspectives for physicians and medical staff in allowing
them to manipulate images from their workstations. The OSIRIS
<http://expasy.hcuge.ch/www/UIN/osiris.html> software has been designed
for the manipulation and analysis of digital medical images stored in a
standard format called PAPYRUS
<http://expasy.hcuge.ch/www/UIN/papyrus.html>. These images can be
retrieved from the PACS database through the ISIS software.

University of Indiana
---------------------
<http://foyt.indyrad.iupui.edu/HomePage.html>. Radiology Home Page -
Indiana University. Many CT and MRI-series transferred to gif-format in
<http://foyt.indyrad.iupui.edu/medres/iurad2.html>.

medical.resources
-----------------
<ftp://ftp.sura.net/pub/nic/> [medical.resources.xx-xx]. It's updated it
4-5 times a year. For the most recent files telnet or Gopher to KUFacts,
(ukanaix.cc.ukans.edu) and login as kufacts. Look under Internet ToolBox
>>> Internet Health Science Resources.

MRI MonkeyHead dataset
----------------------
<ftp://ftp.nc.nihon-u.ac.jp/pub/data/MRIMonkeyHead/>. Image sets of a
brain map on Japanese monkey by using MRI are available which were created
by Dr.Masato Taira of Department of Physiology, Nihon University, School
of Medicine. These image set were anatomically correct. Please use them
for a research on physiology or medical image processing.

NASA Reconstruction Home Page
-----------------------------
<http://biocomp.arc.nasa.gov/3dreconstruction>

>Kevin Montgomery: I'd like to point everyone to a new World-Wide Web
server devoted to 3D reconstruction that I've recently created. It
contains information about software packages (commercial and
researchware), data sets (including pointers to image formats), output
(images/movies), references (bibliographical), and pointers to other
locations over the Internet that contain relevant information (Web, FTP,
news, etc) for CT, MRI, confocal, and serial-section reconstruction. If
you have any additions or modifications to the information contained
therein, please send email to <mailto:[e-mail]> and please
format any new information in HTML if at all possible. This service is
being provided by the NASA Ames Biocomputation Center in order to aid
other researchers in the field and we hope you find the pages useful,
informative, and enjoyable!

NASA URLs
---------
<http://www.nas.nasa.gov/RNR/Visualization/annotatedURLs.html>. A great
resource for general information about visualization related weblets.

National Library of Medicine Visible Human Project
--------------------------------------------------
>Michael J Ackerman: The initial aim of the Visible Human Project is to
create a digital image data set of a complete human male and female
cadaver in MRI, CT and anatomical modes.

The imaging of the male cadaver has been completed. The data set consists
of axial MRI images of the head and neck taken at 5 mm intervals and
longitudinal sections of the rest of the body also at 5 mm intervals. The
MRI images are 256 pixel by 256 pixel resolution. Each pixel has 12 bits
of grey tone resolution.

The CT data consists of axial CT scans of the head and neck taken at 1 mm
intervals at a resolution of 512 pixels by 512 pixels where each pixel is
made up of 12 bits of grey tone. The axial anatomical images are 2048
pixels by 1216 pixels where each pixel is defined by 24 bits of color,
about 7.5 megabytes. The anatomical cross-sections are at 1 mm intervals
and coincide with the CT axial images. About 1880 cross-sections were
obtained from the male cadaver for each mode, i.e., CT and anatomy.

Please remember that this is a "raw" data set. No rendering tools will be
provided with the initial distribution. The data format is a
non-interlaced RGB format read by most advanced rendering systems. Each
anatomical cross-section is over 6 megabytes in size.

The License Agreement for use of the male Visible Human Project data set
is now available. It can be retrieved from NLM's gopher site,
<gopher://gopher.nlm.nih.gov>. The agreement will be found in the section
entitled Visible Human Project as a text file and as a downloadable
WordPerfect file. It is also available from NLM's FTP site,
<ftp://nlmpubs.nlm.nih.gov>. The agreement will be found in section
"visible" as a WordPerfect file, "vhpagree.wp", or as a text file,
"vhpagree.txt". Please make two copies of the agreement and have both
copies signed as originals by your appropriate officials. The agreement
requires that you include a statement explaining your intended use of the
data set. Send both signed copies of the agreement and the statement of
how you intend to use the data set to me at:

Dr. Michael J. Ackerman, Visible Human Project, National Library of
Medicine, 8600 Rockville Pike, Bethesda, MD 20894.

We will have the agreement signed here at NLM and one of the originals
will be returned to you. At that time you will be sent your account and
password to the Visible Human Project FTP site if you wish to download all
or part of the data set via the internet and information on where you may
purchase the data set on 8 mm or 4 mm DAT tape. The data set will be
distributed on 6 DAT tapes in UNIX TAR format corresponding to 6 body
regions. At this time each tape is estimated to cost $150. A 7th,
"sample", tape which contains the entire body at 1 cm increments will also
be available. The set of 7 tapes should cost $1,000. The size of the data
set is about 14 gigabytes.

Sample images are also available at this time via the NLMPubs FTP site.
Six full color anatomical images and an explanatory README file can be
found in <ftp://nlmpubs.nlm.nih.gov/visible/bitmaps/color24/> as "*.raw".
Please be careful as each of these images is over 6 megabytes in size. Ten
CT scan images and an explanatory README file can be found in
<ftp://nlmpubs.nlm.nih.gov/visible/bitmaps/ct/> as "*.fre" (5 images
captured while the cadaver was fresh) and "*.fro" (5 images captured after
the cadaver was frozen). Six MRI scan images and an explanatory README
file can be found in <ftp://nlmpubs.nlm.nih.gov/visible/bitmaps/mri/> as
"*.ti".

The data set from the female cadaver will have the same characteristics as
the male cadaver with one exception. The axial anatomical images will be
obtained at 0.33 mm intervals instead of 1.0 mm intervals. This will
result in over 5,000 anatomical images. The data set is expected to be
about 40 gigabytes in size. Distribution is anticipated during the Summer
of 1995. We are decreasing the spacing in the "Z" direction to 0.33 mm in
order to match the pixel spacing in the "XY" plane which is 0.33 mm. This
will enable developers who are interested in three-dimensional
reconstructions to work with cubic voxels.

If you have further questions, please don't hesitate to ask. Your interest
in NLM's Visible Human Project is greatly appreciated.

Contact: Michael J. Ackerman, Ph.D., Project Officer,
<mailto:[e-mail]>.

<http://www.nlm.nih.gov/extramural_research.dir/visible_human.html>

<gopher://public.nlm.nih.gov/11/visible>

NMR software
------------
<gopher://gopher.nmrfam.wisc.edu>

Penn State University
---------------------
<http://www.xray.hmc.psu.edu/>

Electronic versions of the ACR-NEMA DICOM standard in the directories:
<ftp://ftp.xray.hmc.psu.edu/dicom_docs/dicom_3.0/frame/> framemaker
<ftp://ftp.xray.hmc.psu.edu/dicom_docs/dicom_3.0/word_hqx/> Macintosh Word
<ftp://ftp.xray.hmc.psu.edu/dicom_docs/dicom_3.0/postscript/> postscript

There is also a user conformance profile for DICOM 3.0 written by dr. Fred
Prior of Penn State in the files:
<ftp://ftp.xray.hmc.psu.edu/dicom_docs/UCP.ps>
<ftp://ftp.xray.hmc.psu.edu/dicom_docs/UCP.ps.Z>
<ftp://ftp.xray.hmc.psu.edu/dicom_docs/UCP.txt>
<ftp://ftp.xray.hmc.psu.edu/dicom_docs/UCP.txt.X>

We also offer access to the Mallinckrodt Institute of Radiology's DICOM
implementation in the directory:
<ftp://ftp.xray.hmc.psu.edu/dicom_software/Mallinckrodt/>

We are also pleased to announce the availability of the European
CEN/TC251/WG4 version of DICOM software in the directory:
<ftp://ftp.xray.hmc.psu.edu/dicom_software/European/>

Clinical xray mammography images which have been digitized on a Lumysis
scanner. The images were obtained from screening sessions, and the breasts
were determined normal in the screening. No subsequent images were taken,
so these images are not verified as normal based on followup examinations.
There are two digitizations of each image, one at high resolution, and one
at low resolution. There are 12 significant bits/pixel, though each pixel
is stored at 16 bits/pixel in order minimize the complexity of reading the
images. The high resolution image was digitized at 50 micron spot size and
the low resolution image was digitized at 100 micron spot size. The 50
micron images are 3048x4318 pixels (6"x8.5") and the 400 micron images are
381x540 pixels. Patient information was not scanned. The high and low
bytes will be reversed for Macintoshes and Sun Workstations. For example,
when reading an image into NIH Image for the Macintosh, the Byte Reversed
toggle must be selected. The pixel value is nominally pv = 1000*(Optical
Density), with a useful range of 0.1 to 3.0. Each image begins with a 2048
byte Lumysis header. The only really useful information the header
contains is the number of pixels in the x and y directions, and the spot
size.
<ftp://ftp.xray.hmc.psu.edu/mammo/>

Interactive Computing Resource of the Magnetic Metabolic Research Resource
Computing Center: <ftp://nmrsg.biophys.upenn.edu/pub/>. New developments
in this FTP site are documented in the electronic newsletter, WavePacket.
You can subscribe to WavePacket by sending e-mail to
<mailto:[e-mail]> and putting "subscribe" in the
first line of the text.
/pub/data         interesting MR and optical data sets
/pub/education    technical documents of educational nature
/pub/programs     programs for MR related analysis, display, ...
/pub/sequences    pulse sequence programs for various MR systems
/pub/wavepacket   editions of WavePacket, the electronic newsletter

Radiology Teaching Files on the Web
-----------------------------------
There are many sites offering Radiology teaching files on the Web.
Unfortunately, each site is completely independent of the others and there
is no common search method to find all the radiology teaching files. As a
temporary solution in time for board review we offer the following list of
Radiology Teaching files by subject. Note that none of the files below
reside here at Penn State, rather we have visited each site in our list of
colleges and universities, extracted those that offer teaching files and
organized the links by subject rather than by institution. Enjoy! (perhaps
someday we can organize a registry service for teaching files).

Anyone know of any radiology related teaching files to add??? If yes,
please send pointers to David S. Channin, <mailto:[e-mail]>.

<http://www.xray.hmc.psu.edu/public/tf.html>

RSNA
----
<http://www.rsna.org/>
<http://www.rsna.org/edu/publications/pub.html> includes recent table of
contents of Radiology and Radiographics.

<ftp://rsna.org/pub/>

RSNA maintains a public mailing list about DICOM. To subscribe send mail
with the command "add" in the text of your message to
<mailto:[e-mail]>. To send mail to the group, mail to
<mailto:[e-mail]>.

Scientific visualization home page
----------------------------------
<http://web.msi.umn.edu/WWW/SciVis/umnscivis.html>. Scientific
visualization home page. Contact: <mailto:[e-mail]>.

sci.image.processing archive
----------------------------
<ftp://ruby.oce.orst.edu/pub/sci.image.processing/> or
gopher://skyking.oce.orst.edu, on port 71 "Information from the Coastal
Imaging Lab". sci.image.processing archive.

UMDS Image Processing Group
---------------------------

<http://www-ipg.umds.ac.uk/>. The Image Processing Group [IPG] is an
interdisciplinary research group based at UMDS, the United Medical &
Dental Schools of Guy's and St Thomas' Hospitals.

We are developing image processing and computer vision techniques, with a
particular emphasis on medical applications. This WWW has information
about The Image Processing Group, including staff, publications and
project pages, the IPG's Medical Image Archive, the IPG's Teaching
Archive, Biomedical Imaging Conference Database, Radiological Sciences
Services, Radiological Sciences newsletter, Job Vacancies. Includes an
example of combining MR, CT, and MR angiographic images for planning skull
base surgery.

<ftp://boris.umds.ac.uk/pub/images/>

ct01_raw.Z    6073k    24.5.1994
mr01_raw.Z    3197k    24.5.1994
mr02_raw.Z    6099k    24.5.1994
mr03_raw.Z    4862k    30.8.1994
pet02_raw.Z    1097k    24.5.1994

WWW news.answers archive
------------------------
>Henk Penning <mailto:[e-mail]>: Take a look at our new WWW
news.answers archive at <http://www.cs.ruu.nl/cgi-bin/faqwais>. The
archive supports full-text search by keyword (WAIS), access by
archive-name/subject and access by news.group.

Each article contains a link to the archive-name directory where the
article resides, and link(s) to overviews of the newsgroup(s) where the
article was posted.

This means that when you do a keyword search, you can jump from any
matching article to other related articles in the same archive-name
directory and to other articles posted in the same news.group(s).

Articles are left intact, but things that look like URLs are activated
(made selectable).

The archive is generated daily from our news.answers ftp-archive by two
small Perl programs. It takes about 1.5 hours (sparc-2).

xmovie.skull.tar.Z
------------------
<ftp://camelot.usc.edu/pub/images/xmovie.skull.tar.Z>

>AJ Annala: If you might be interested in viewing a 36 frame animation
loop of a series of views taken at 10 degree intervals around a 3D stack
of 113 256x256 pixel CAT scan images of a human head then acquire the
file, compile the xmovie program on your system (modify the makefile to
use your local C compiler, X11 lib, and socket lib) -- then enter 'xmovie
-height 256 -width 256 skull*' to view the animation loop on your X11R4 or
X11R5 display station. If you have the ability to select your local X11
screen resolution (SUN's can't do this -- but most i386 unix systems can
select their display resolution) you may enlarge the image by selecting an
X11 display resolution of 640 x 480 x 8 bit color. Otherwise you will be
viewing a 256 x 256 pixel animated image within whatever screen resolution
is the default on your system.

Newsgroups of (possible) interest
---------------------------------
alt.3d
alt.graphics
alt.graphics.pixutils
alt.image.medical
alt.sci.nmr
bionet.neuroscience
bionet.software
comp.graphics
comp.graphics.algorithms
comp.graphics.avs
comp.graphics.data-explorer
comp.graphics.explorer
comp.graphics.visualization
comp.lang.idl-pvwave
comp.protocols.dicom
comp.soft-sys.khoros
comp.soft-sys.wavefront
comp.sys.mac.graphics
comp.sys.sgi.graphics
sci.answers
sci.data.formats
sci.image.processing
sci.med
sci.med.physics
sci.med.radiology
sci.med.telemedicine
sci.med.vision
sci.techniques.mag-resonance
sci.techniques.microscopy

DICOM info, software and example image files
============================================

ACR Nema acquisition Plug-In for Mac
------------------------------------
ACR Nema acquisition Plug-In (Macintosh) for Photoshop and NIH-Image:
<ftp://zippy.nimh.nih.gov/pub/nih-image/plug-ins/ACRNema.hqx> or:
<ftp://ftp.funet.fi/pub/mac/info-mac/grf/util/photoshop-acr-nema.hqx>

David Clunie's dicom tools
--------------------------
<ftp://ftp.rahul.net/pub/dclunie/dicom3tools_0.08.tar.gz>. Tools and
libraries for handling files of DICOM 3 attributes, and conversion of
proprietary formats to DICOM 3. Can handle older ACR/NEMA format data, and
some proprietary versions of that such as SPI. Also handles Part 10
Metaheaders. VERY limited X display capability.

Proprietary image conversions from General Electric CT 9800, General
Electric CT High Speed Advantage (Genesis), General Electric MR Signa
3X/4X, General Electric MR Signa 5X (Genesis), General Electric CT Sytec,
Siemens Somatom CT DR family, Siemens Somatom Plus family (SPI version of
ACR/NEMA), Siemens Magnetom Impact (SPI version of ACR/NEMA), Siemens
Magnetom SP (SPI version of ACR/NEMA), Philips Gyroscan MR S5 (native &
exported SPI(ANI)).

Image format support: DICOM 3 offline file format as per draft Part 10,
Parsing/validating DICOM 3 data sets as modules and IODs, Pbmplus extended
16 bit raw format, Raw binary images.

Archive retrieval from General purpose 9-track and DAT file extraction,
General Electric CT 9800 9-track, General Electric Genesis DAT, Philips
Gyroscan MR S5 native format (ANSI format tapes).

Miscellaneous image format utilities: Dump (octal / hex / decimal / byte /
short / long / ieee float), Patch, Swap bytes, Word to byte shift, Vax VMS
DUMP output to binary (poor man's uudecode).

Frequent Answers: Yes, many of the tools compile on a Mac (Symantec) but
there are no screen based utilites yet. No, I haven't tried compiling on a
PC under DOS or Windows, though it works under Linux if you have enough
memory for the compile. Yes, I will write a viewer for the Mac (and maybe
even Windows if I ever succumb and by a PC), but it is a low priority ...
try the Photoshop ACR/NEMA plugin. No, the tools are not a network DICOM
package, they just convert things to offline file formats (for now). Yes,
if you don't care about dicom (!) but want to translate from a proprietary
format to something else, the tools will help. No I don't know the format
of the GE Genesis optical disks. Yes, I do want to hear from ANYONE who
knows ANYTHING about any medical image format that is not included, or can
provide sample images to reverse engineer.

Comments, criticism and general abuse are greatly appreciated and should
be directed to <mailto:[e-mail]>.

Dr Razz
-------
[look under software packages]

General Electric
----------------
[look under some medical sites]

ImportACCESS Plug-In for Mac
----------------------------
Designed Access of Chicago, IL announces the introduction of ImportACCESS,
a commercially available product for importing medical and scientific
files directly into a wide variety of software packages previously
incapable of handling such data.

Written as an Adobe Photoshop plug-in for the Macintosh line of personal
computers, ImportACCESS brings CT, MR, SPECT, PET, and other forms of
digitally collected data to the desktop, regardless of the format in which
the data has been saved. By supporting most raw data formats as well as
evolving standards such as ACR-NEMA 2.0, DICOM 3.0, Interfile 3.3, and
Papyrus, ImportACCESS provides both backward and forward compatibility for
accessing imaging files. It allows software packages as Adobe Photoshop,
NIH-Image, and over twenty other programs to be used as viewboxs for
visualization, presentation, and creation of various forms of B/W and
color output.

Support is included for window level manipulation, template creation for
formatting single and multi-slice files, and tools for consistent
production of publication quality output. An integrated format editor with
real-time file preview is used for fixed file format creation, and a
drop-in reader interface is provided for tagged formats such as DICOM 3.0,
and other newly emerging industry standards.

ImportACCESS works with any Macintosh running System 7.0 or later. The
programs costs US$449, with external code readers for DICOM 3.0, ACR-NEMA
2.0, Interfile 3.3, and Papyrus costing US$75 per reader.

Contact: Hugh Lyshkow, Chief Technical Officer, Designed Access, 702
Wrightwood Avenue, Chicago, IL 60614, Phone: [phone], FAX: (312)
472-8834, <mailto:[e-mail]>.

NIH-Image
---------
[look under software packages]

NIH-Image Macro to import ACR-NEMA 2.0 and  DICOM 3.0 images by Tunc
Iyriboz <mailto:[e-mail]>: >I tried it with a various
ACR-NEMA and some DICOM 3.0 images, coming from Internet archives, from
various manufacturers like Elscint, Picker, and third-party developers
like Evergreen Technologies. I had problems importing KODAK PDS 2.5
storage files, using an inversed byte order and separate header file. It
certainly should not work with many other semi-proritary format files. It
doesn't work yet either with part 10 complient DICOM 3.0 files
unfortunately. I am working on part 10 compatible images from GE.
<ftp://zippy.nimh.nih.gov/pub/nih-image/user-macros/import_acr_nema.txt>

PAPYRUS
-------
PAPYRUS 3.0 file format is based on the new DICOM 3.0 Standard. Although
version 3.0 of PAPYRUS is absolutely compliant with DICOM Part 10, it does
reinforce a few rules. It provides a way to group several images in a file
in a "PAPYRUS-File object". There are certainly other ways to regroup
images in a single file while remaining compliant with DICOM part 10
specifications . PAPYRUS is also restricted to keep images of a same
patient and of a same series together. PAPYRUS is not intended to stores
images from different patients (teaching files for example) or from
different modalities or even from different acquisition series. These
restrictive rules allow a better file management in a PACS environment.
PAPYRUS 3.0 is presented as a "profile" of DICOM possible implementations.
While PAPYRUS specification focuses on the image files it still relies on
the directory structure defined in Part 10 to keep these files linked
together on a given storage media (DICOMDIR file).

<http://expasy.hcuge.ch/www/UIN/papyrus.html>

<ftp://expasy.hcuge.ch/pub/Osiris/Images/>

Penn State University
---------------------
[look under some medical sites]

RSNA
----
[look under some medical sites]

Example DICOM image files at wuerlim.wustl.edu
----------------------------------------------
<ftp://wuerlim.wustl.edu/pub/dicom/images/version3/>. This directory
contains example DICOM image files that were created for the DICOM
demonstration held at the annual meeting of the Radiological Society of
North America, December, 1993. These images are stored in Little-Endian
byte order as a stream of bytes. This file format is not the format that
is defined in Part 10 of the Standard. These image files are readable with
the MIR CTN software. Each tar file contains one directory with images
that consitute a study. Some of the images may not have all fields filled
in correctly. Specifically, the geometry information is probably incorrect
(image position, image orientation). Any values that correspond to
acquisition parameters are also suspect.

cr1.tar    4192k
cr2.tar    4112k
cr3.tar    3480k
ct1.tar   11984k
ct2.tar    4328k
ct3.tar    5024k
ct4.tar   12632k
mr1.tar   11760k
mr2.tar    5512k
mr3.tar   10128k
mr4.tar    8336k
us1.tar      96k
us2.tar      48k
us3.tar     144k
xray1.tar  5568k
xray2.tar 10960k
xray3.tar  5208k
xray4.tar  5544k

Other interesting FAQs
======================

Medical Image Format FAQ
------------------------
From: [e-mail] (David A. Clunie)
Newsgroups: alt.image.medical,comp.protocols.dicom,sci.data.formats,
            alt.answers,comp.answers,sci.answers,news.answers
Subject: Medical Image Format FAQ
Summary: This posting contains answers to the most Frequently Asked
         Question on alt.image.medical - how do I convert from image
         format X from vendor Y to something I can use ? In addition
         it contains information about various standard formats.

<ftp://rtfm.mit.edu
/pub/usenet-by-group/news.answers/medical-image-faq/> [part1-3]

<ftp://ftp.rahul.net/pub/dclunie/medical-image-faq/> Text, html or
postscript versions.

<http://www.rahul.net/dclunie/medical-image-faq/html/>

comp.graphics.vis FAQ
---------------------
From: [e-mail] (Eugene N. Miya)
Newsgroups: comp.graphics.visualization
Subject: comp.graphics.vis FAQ c.g.v.FAQ
Organization: NASA Ames Research Center, Moffett Field, CA

comp.compression Frequently Asked Questions
-------------------------------------------
From: [e-mail] (Jean-loup Gailly)
Newsgroups: comp.compression,comp.compression.research,
            news.answers,comp.answers
Subject: comp.compression Frequently Asked Questions

<ftp://rtfm.mit.edu
/pub/usenet-by-group/news.answers/compression-faq/> [part1-3]

JPEG image compression: Frequently Asked Questions
--------------------------------------------------
From: [e-mail] (Tom Lane)
Subject: JPEG image compression: Frequently Asked Questions
Newsgroups: comp.graphics,alt.graphics.pixutils,
            alt.binaries.pictures.utilities,
            alt.binaries.pictures.d,
            alt.binaries.pictures.erotica.d,comp.answers,
            alt.answers,news.answers
Summary: Useful info about JPEG (JPG) image files and programs

<ftp://rtfm.mit.edu/pub/usenet-by-group/news.answers/jpeg-faq>

comp.graphics Frequently Asked Questions
----------------------------------------
From: [e-mail] (John T. Grieggs)
Newsgroups: comp.graphics,comp.answers,news.answers
Subject: comp.graphics Frequently Asked Questions (FAQ)

<ftp://rtfm.mit.edu/pub/usenet-by-group/news.answers/graphics/faq>

Computer Graphics Resource Listing
----------------------------------
From: [e-mail] (Nick C. Fotis)
Subject: Computer Graphics Resource Listing
Newsgroups: comp.graphics,comp.answers,news.answers

<ftp://rtfm.mit.edu
/pub/usenet-by-group/news.answers/graphics/resources-list/> [part1-6]

Macintosh Image Processing Information Sources (FAQ)
----------------------------------------------------
From: [e-mail] (Edward J. Huff)
Newsgroups:
sci.image.processing,comp.sys.mac.scitech,sci.answers,news.answers
Subject: Macintosh Image Processing Information Sources (FAQ)
Summary: Macintosh Image Processing Information available via gopher, FTP,
         Usenet, e-mail, telephone, and snail mail.
X-Last-Updated: 1993/10/26

<ftp://rtfm.mit.edu
/pub/usenet-by-group/news.answers/image-processing/Macintosh>

Scientific Data Format Information FAQ
--------------------------------------
From: [e-mail] (Ilana Stern)
Newsgroups: sci.data.formats,news.answers,sci.answers
Subject: Scientific Data Format Information FAQ

<ftp://rtfm.mit.edu/pub/usenet-by-group/news.answers/sci-data-formats>

-end of med-volviz-faq-95-03-